Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks465 from public data
- StackGAN: Text to Photo-Realistic Image Synthesis with Stacked Generative Adversarial Networks2,975
This paper proposes Stacked Generative Adversarial Networks (StackGAN) to generate 256 photo-realistic images conditioned on text descriptions and introduces a novel Conditioning Augmentation technique that encourages smoothness in the latent conditioning manifold.
- StackGAN++: Realistic Image Synthesis with Stacked Generative Adversarial Networks1,198
Extensive experiments demonstrate that the proposed stacked generative adversarial networks significantly outperform other state-of-the-art methods in generating photo-realistic images.
- Engineering and Algorithm Design for an Image Processing API: A Technical Report on ITK - the Insight Toolkit745
The detailed planning and execution of the Insight Toolkit (ITK), an application programmers interface (API) for the segmentation and registration of medical image data, is presented.
- Learning with structured sparsity571
This paper investigates a new learning formulation called structured sparsity, which is a natural extension of the standard sparsity concept in statistical learning and compressive sensing by allowing arbitrary structures on the feature set, which generalizes the group sparsity idea.
- SVDiff: Compact Parameter Space for Diffusion Fine-Tuning429
The proposed SVDiff method involves fine-tuning the singular values of the weight matrices, leading to a compact and efficient parameter space that reduces the risk of overfitting and language-drifting, making it more practical for real-world applications.
- 385
- An anthropometric face model using variational techniques377
A system that automatically generates varied geometric models of human faces according to anthropometric statistics for likely face measurements in a population using Variational modeling to find a smooth surface that satisfies constraints while using a prototype shape as a reference.
- A Framework for Recognizing the Simultaneous Aspects of American Sign Language364
This paper presents a novel framework to ASL recognition that aspires to being a solution to the scalability problems, based on breaking down the signs into their phonemes and modeling them with parallel hidden Markov models.
- Optical Flow Constraints on Deformable Models with Applications to Face Tracking360
A dynamic system incorporating flow as a hard constraint is derived and solved, producing a model-based least-squares optical flow solution that ensures the constraint remains satisfied when combined with edge information, which helps combat tracking error accumulation.
- 347
- Oriented Object Detection in Aerial Images with Box Boundary-Aware Vectors329
This work extends the horizontal keypoint-based object detector to the oriented object detection task, and shows that learning the box boundary-aware vectors is superior to directly predicting the width, height, and angle of an oriented bounding box, as adopted in the baseline method.
- 301
- 259
- Autism: the micro-movement perspective257
The methods address the heterogeneity of ASD with a personalized approach grounded in the inherent sensory-motor abilities that the individual has already developed, based on the non-stationary stochastic patterns of minute fluctuations inherent to the authors' natural actions.
- Automatic Liver Segmentation Using an Adversarial Image-to-Image Network249
This paper proposes an automatic and efficient algorithm to segment liver from 3D CT volumes using a deep image-to-image network employing a convolutional encoder-decoder architecture combined with multi-level feature concatenation and deep supervision.
- 249
- Novel method for high-throughput phenotyping of sleep in mice245
The method gives excellent agreement in C57BL/6J male mice with simultaneous assessment of sleep by EEG/EMG recording and has applications in chemical mutagenesis and for studies of molecular changes in brain with sleep/wakefulness.
- 241
- Semantic Amodal Segmentation228
A detailed image annotation that captures information beyond the visible pixels and requires complex reasoning about full scene structure is proposed, and it is shown that the proposed full scene annotation is surprisingly consistent between annotators, including for regions and edges.
- 228
- Jointly Optimize Data Augmentation and Network Training: Adversarial Data Augmentation in Human Pose Estimation226
The key idea is to design a generator that competes against a discriminator that explores weaknesses of the discriminators, while the discriminator learns from hard augmentations to achieve better performance.
- 222
- SINE: SINgle Image Editing with Text-to-Image Diffusion Models220
A novel model-based guidance built upon the classifier-free guidance is proposed so that the knowledge from the model trained on a single image can be distilled into the pre-trained diffusion model, enabling content creation even with one given image.
- 220
- 216
- MotionNet: Joint Perception and Motion Prediction for Autonomous Driving Based on Bird’s Eye View Maps197
Extensive experiments show that the proposed method overall outperforms the state-of-the-arts, including the latest scene-flow- and 3D-object-detection-based methods, indicating the potential value of the proposed methods serving as a backup to the bounding-box-based system, and providing complementary information to the motion planner in autonomous driving.
- Query Specific Fusion for Image Retrieval197
A graph-based query specific fusion approach where multiple retrieval sets are merged and reranked by conducting a link analysis on a fused graph is proposed, capable of adaptively integrating the strengths of the retrieval methods using local or holistic features for different queries without any supervision.
- 197
- A Variational Level Set Approach to Segmentation and Bias Correction of Images with Intensity Inhomogeneity196
A variational level set approach to joint segmentation and bias correction of images with intensity inhomogeneity that is robust to initialization, and therefore allows automatic applications, and has been used for images of various modalities with promising results.
- 187
- CR-GAN: Learning Complete Representations for Multi-view Generation158
The results prove that CR-GAN significantly outperforms state-of-the-art methods, especially when generating from ?
- Deformable segmentation via sparse representation and dictionary learning156
Two strategies to decrease the computational complexity of SSC are designed, making a robust, accurate and efficient deformable segmentation system and improve the overall accuracy.
- Reconstruction-Based Disentanglement for Pose-Invariant Face Recognition150
This paper presents a method for learning a feature representation that is invariant to pose, without requiring extensive pose coverage in training data, and proposes a new feature reconstruction metric learning to explicitly disentangle identity and pose.
- A Recurrent Encoder-Decoder Network for Sequential Face Alignment146
A novel recurrent encoder-decoder network model for real-time video-based face alignment that predicts 2D facial point maps regularized by a regression loss, while uniquely exploiting recurrent learning at both spatial and temporal dimensions is proposed.
- Quantized Densely Connected U-Nets for Efficient Landmark Localization144
This paper proposes quantized densely connected U-Nets for efficient visual landmark localization with order-K dense connectivity to trim off long-distance shortcuts and uses a memory-efficient implementation to significantly boost the training efficiency and investigates an iterative refinement that may slice the model size in half.
- 131
- Deep Animation Video Interpolation in the Wild127
This work formally defines and study the animation video interpolation problem for the first time, and proposes an effective framework, AnimeInterp, with two dedicated modules in a coarse-to-fine manner, which outperforms existing state-of-the-art interpolation methods for animation videos.
- Human gait recognition at sagittal plane127
This paper proposes a novel two-step, model-based approach to gait recognition by employing a five-link biped locomotion human model, and minimizes the size of the data required for recognition compared to model-free algorithms.
- Three-dimensional motion reconstruction and analysis of the right ventricle using tagged MRI127
A method for reconstructing the 3D motion of the RV by fitting a deformable model to tag and contour data extracted from multiview tagged magnetic resonance images, which accommodates the geometrically complex RV.
- 121
- Automating gait generation119
ElevWalker, a new low-level gait generator based on sagittal elevation angles, which allows curved locomotion - walking along a curved path - to be created easily; ElevInterp, which uses a new inverse motion interpolation algorithm to handle uneven terrain locomotion; and MetaGait, a high-level control module which allows an animator to control a figure's walking simply by specifying a path are presented.
- Attentive neural cell instance segmentation118
This paper proposes an attentive instance segmentation method that accurately predicts the bounding box of each cell as well as its segmentation mask simultaneously simultaneously, and employs the attention mechanism in both detection and segmentation modules to focus the model on the useful features.
- Handshapes and Movements: Multiple-Channel American Sign Language Recognition114
A framework for recognizing American Sign Language (ASL) and the main challenges in developing scalable recognition systems are to devise the basic building blocks from which to build up the signs, and to handle simultaneous events.
- Global Matching with Overlapping Attention for Optical Flow Estimation111
This paper introduces a simple but effective global matching step before the direct regression and develops a learning-based matching-optimization framework, namely GMFlowNet, which outperforms RAFT, the most popular optimization-only method, by a large margin and achieves state-of-the-art performance on standard benchmarks.
- 111
- Automatic Vertebra Labeling in Large-Scale 3D CT Using Deep Image-to-Image Network with Message Passing and Sparsity Regularization110
This paper proposes an automatic and fast algorithm to localize and label the vertebra centroids in 3D CT volumes and outperforms other state-of-the-art methods in terms of localization accuracy.
- Optical computer recognition of facial expressions associated with stress induced by performance demands.109
Results suggest that an OCR algorithm using mouth and eyebrow regions has the potential to discriminate high- from low-stressor performance bouts in 75-88% of subjects, and the validity of the workload paradigm to induce differential levels of stress in facial expressions was established.
- Learning to Forecast and Refine Residual Motion for Image-to-Video Generation106
This work trains networks to learn residual motion between the current and future frames, which avoids learning motion-irrelevant details and proposes a two-stage generation framework where videos are generated from structures and then refined by temporal signals.
- 105
- Boosting encoded dynamic features for facial expression recognition104
This paper designs the dynamic haar-like features to represent the temporal variations of facial appearance and further encode the dynamic features into the binary pattern features, which are useful to construct weak classifiers for boosting learning.
- ProxEdit: Improving Tuning-Free Real Image Editing with Proximal Guidance103
This work enhances NPI with a regularization term and inversion guidance, which reduces artifacts while capitalizing on its training-free nature, and extends the concepts to incorporate mutual self-attention control, enabling geometry and layout alterations in the editing process.
- 103
- Background Subtraction Using Low Rank and Group Sparsity Constraints102
A unified and robust framework to effectively handle diverse types of videos, e.g., videos from stationary or moving cameras, is proposed using both low rank and group sparsity constraints.
- MRI Reconstruction Via Cascaded Channel-Wise Attention Network101
A new model called MRI Cascaded Channel-wise Attention Network (MICCAN), highlighted by three components, is proposed, able to attend to salient information by filtering irrelevant features and also concentrate on high-frequency information by enforcing low-frequency Information bypassed to the final output on the MRI reconstruction with low undersampled k-space data.
- ASSD: Attentive single shot multibox detector98
The proposed network, termed ASSD, builds feature relations in the spatial space of the feature map with the global relation information, and learns to highlight useful regions on the feature maps while suppressing the irrelevant information, thereby providing reliable guidance for object detection.
- 96
- 95
- 94
- FocusNetv2: Imbalanced large and small organ segmentation with adversarial shape constraint for head and neck CT images93
A novel two-stage deep neural network, FocusNetv2, is proposed to solve this challenging problem by automatically locating, ROI-pooling, and segmenting small organs with specifically designed small-organ localization and segmentation sub-networks while maintaining the accuracy of large organ segmentation.
- Three-dimensional systolic kinematics of the right ventricle92
The combination of tagged MRI and three-dimensional (3-D) image-processing techniques was used to reconstruct 3-D RV-LV motion and deformation and it was found that the RV exhibited a similar twisting pattern, a more longitudinal strain pattern, and a greater amount of displacement than the LV.
- Temporal spectral residual91
A new method Temporal Spectral Residual on video slices along X-T and Y-T planes, which can automatically separate foreground motion objects from backgrounds, also with the help of threshold selection and voting schemes is proposed.
- 89
- Animation and control of breaking waves88
A novel fluid animation control approach that provides a significantly faster method for obtaining the full 3D breaking wave evolution compared to starting the simulation at an early stage and using solely the 3D Navier-Stokes equations.
- 86
- 85
- 76
- Knowledge As Priors: Cross-Modal Knowledge Generalization for Datasets Without Superior Knowledge74
This paper proposes a novel scheme to train the Student in a Target dataset where the Teacher is unavailable by modeling knowledge as priors on parameters of the Student by generalizing the distilled cross-modal knowledge learned from a Source dataset.
- The Hidden Life of Tokens: Reducing Hallucination of Large Vision-Language Models via Visual Information Steering73
This paper proposes VISTA (Visual Information Steering with Token-logit Augmentation), a training-free inference-time intervention framework that reduces hallucination while promoting genuine information in activation space and leveraging early layer activations to promote semantically meaningful decoding.
- 73
- Learning Trailer Moments in Full-Length Movies with Co-Contrastive Attention67
A novel ranking network is introduced that utilizes the Co-Attention between movies and trailers as guidance to generate the training pairs, where the moments highly corrected with trailers are expected to be scored higher than the uncorrelated moments.
- A review of motion analysis methods for human Nonverbal Communication Computing67
The concept and applications of Nonverbal Communication Computing are introduced and some of the motion analysis methods employed in this area are reviewed, which include face tracking, expression recognition, body reconstruction, and group activity analysis.
- Handshapes and movements: Multiple-channel ASL recognition66
This paper loosely follow the Movement-Hold model to devise a breakdown of the signs into their constituent phonemes, which provide the fundamental building blocks, and shows how to integrate the handshape into this breakdown, and discusses what handshape representation works best.
- 65
- 65
- Facial expression recognition using encoded dynamic features65
A novel framework for video-based facial expression recognition, which can handle the data with various time resolution including a single frame, is proposed, which uses the haar-like features to represent facial appearance and performs K-Means clustering on the facial appearance features to explore the intrinsic temporal patterns of each expression.
- Algorithmic issues in modeling motion65
The aim of the article is to review current approaches to modeling motion together with related data structures and algorithms, and to summarize the challenges that lie ahead in producing a more unified theory of motion representation that would be useful across several disciplines.
- Revisiting Multimodal Representation in Contrastive Learning: From Patch and Token Embeddings to Finite Discrete Tokens63
It is demonstrated that using FDT representations in CLIP-style models improves cross-modal alignment and performance in visual recognition and vision-language downstream tasks, and it is shown that the method can learn more comprehensive representations, and the learned FDT capture meaningful cross- modal correspondence, ranging from objects to actions and attributes.
- 63
- Automated 3D segmentation using deformable models and fuzzy affinity63
An algorithm for segmenting objects with closed, non-intersecting boundaries, such as the heart and the lungs, that is independent of the imaging modality used is developed, which automates and significantly improves the object boundary detection results.
- Simulation of two‐phase flow with sub‐scale droplet and bubble effects62
A new Eulerian‐Lagrangian method for physics‐based simulation of fluid flow, which includes automatic generation of sub‐scale spray and bubbles, is presented, which is quite powerful in its applications.
- Consensus of Regression for Occlusion-Robust Facial Feature Localization61
This work proposes an occlusion-robust regression method, which is trained to estimate facial feature locations under the precondition that a particular pre-defined region of the face is occluded, and achieves consistently better results over state-of-the-art methods for facial landmark localization and Occlusion detection.
- 60
- Hypergraph with sampling for image retrieval59
A new transductive learning framework for image retrieval is proposed, in which images are taken as vertices in a weighted hypergraph and the task of image search is formulated as the problem of hypergraph ranking.
- Score-Guided Diffusion for 3D Human Recovery58
By guiding its denoising process with a task-specific score, ScoreHMR effectively solves inverse problems for various applications without the need for retraining the task-agnostic diffusion model.
- DeepTag: An Unsupervised Deep Learning Method for Motion Tracking on Cardiac Tagging Magnetic Resonance Images55
This paper proposes a novel deep learning-based fully unsupervised method for in vivo motion tracking on t-MRI images and shows that this method is superior to conventional motion tracking methods in terms of landmark tracking accuracy and inference efficiency.
- Multi-scale Cell Instance Segmentation with Keypoint Graph Based Bounding Boxes55
This paper proposes a new box-based cell instance segmentation method that first detects the five pre-defined points of a cell via keypoints detection, then group these points according to a keypoint graph and subsequently extract the bounding box for each cell.
- 55
- 54
- 54
- Motion analysis of the right ventricle from MRI Images54
This paper uses data from tagged MRI images to perform the first 3D motion study of the entire right ventricle to date and builds a biventricular model, focusing on method validation and visualizing clinically useful parameters that describe RV wall motion.
- Optimal Topological Cycles and Their Application in Cardiac Trabeculae Restoration52
A novel method to jointly detect salient topological handles and compute the optimal representations of them is proposed and shown in experiments the significance of this contribution compared with previous standard segmentation methods without topological priors, as well as with previous topological method in which non-optimal representations of topological handle are used.
- 52
- 52
- 52
- 52
- Motion Profiles for Deception Detection Using Visual Cues51
A data-driven, unobtrusive and covert method for automatic deception detection in interrogation interviews from visual cues only, using skin blob analysis together with Active Shape Modeling to extract motion profiles, thus aiding in the discrimination of truthful and deceptive responses.
- A homotopy-based sparse representation for fast and accurate shape prior modeling in liver surgical planning49
This work employed the Sparse Shape Composition (SSC) to represent liver shapes by an optimized sparse combination of shapes in the repository, without any assumptions on parametric distributions of liver shapes, and introduced a homotopy-based method to quickly solve the L1-norm optimization problem in SSC.
- Patch-Based Texture Edges and Segmentation49
A novel technique for extracting texture edges based on the combination of two ideas: the patch-based approach, and non-parametric tests of distributions, which can reliably detect texture edges using only local information and can be combined with parametric deformable models.
- Show Me What and Tell Me How: Video Synthesis via Multimodal Conditioning48
To improve video quality and consistency, a new video token trained with self-learning and an improved mask-prediction algorithm for sampling video tokens are proposed and text augmentation is introduced to improve the robustness of the textual representation and diversity of generated videos.
- Deformable Segmentation via Sparse Shape Representation48
A novel deformable model is proposed for robust segmentation in the presence of weak/misleading appearance cues and achieves better performance in both studies.
- 47
- Social ODE: Multi-agent Trajectory Forecasting with Neural Ordinary Differential Equations46
This work proposes Social ODE which explicitly models temporal agent dynamics and agent interactions and leverages Neural ODEs to model continuous temporal dynamics, and incorporates distance, interaction intensity, and aggressiveness estimation into agent interaction modeling in latent space.
- Bodypart Recognition Using Multi-stage Deep Learning46
A multi-stage deep learning framework that aims to discover the local regions that are discriminative to the bodypart recognition, and learn a bodypart identifier based on these local regions, which achieves better performances than state-of-the-art approaches, including the standard CNN.
- Deformable models with sparsity constraints for cardiac motion analysis46
This paper introduces a new family of deformable models that are inspired from the compressed sensing, a technique for accurate signal reconstruction by harnessing some sparseness priors, and employs sparsity constraints to handle the outliers or gross errors.
- Monitoring of facial stress during space flight: Optical computer recognition combining discriminative and generative methods46
A comprehensive method to detect stress from facial image sequences by using a model-based tracker and consisting of generative and discriminative parts that yield results superior to using either approach individually is proposed.
- Open science – combining open data and open source software: Medical image analysis with the Insight Toolkit46
A trend in the health sciences is emerging that supports collaborative research and development with a strong focus on clinical practice, early diagnosis, and the discovery of new treatments.
- American sign language recognition: reducing the complexity of the task with phoneme-based modeling and parallel hidden markov models46
This thesis presents a framework for recognizing American Sign Language (ASL) from 3D data, and develops a PaHMM recognition algorithm specifically geared toward the properties of sign languages, and poses parallel hidden Markov models (PaHMMs) as an extension to conventional hiddenMarkov models.
- Scalable histopathological image analysis via supervised hashing with multiple features45
This work exploits joint kernel-based supervised hashing (JKSH) to integrate complementary features in a hashing framework to bridge the semantic gap between low-level features and high-level diagnosis in histopathological image analysis methods.
- 44
- 43
- An efficient conditional random field approach for automatic and interactive neuron segmentation43
An interactive segmentation framework which only selects locations where the model is uncertain for a user to proofread is developed and modifies the merging tree and thus improves the segmentation globally.
- 43
- Steering Rectified Flow Models in the Vector Field for Controlled Image Generation42
This work develops a theoretical and empirical understanding of the vector field dynamics of RFMs in efficiently guiding the denoising trajectory and proposes FlowChef, which leverages the vector field to steer the denoising trajectory for controlled image generation tasks, facilitated by gradient skipping.
- 42
- 41
- Image Segmentation Based on the Integration of Markov Random Fields and Deformable Models41
This paper makes use of Markov Random Field theory to build a Gibbs Prior model of medical images with arbitrary initial parameters to estimate the boundary of organs with low signal to noise ratio (SNR).
- Taming Self-Training for Open-Vocabulary Object Detection40
This work proposes SAS-Det, a split-and-fusion (SAF) head that splits a standard detection into an open-branch and a closed-branch and introduces a periodic update strategy to decrease the number of up-dates to the teacher, thereby decreasing the frequency of changes in PL distributions, which stabilizes the training process.
- Diffusion Guided Domain Adaptation of Image Generators40
It is shown that the classifier-free guidance can be leveraged as a critic and enable generators to distill knowledge from large-scale text-to-image diffusion models to adapt a GAN generator to another domain.
- 39
- 39
- Sharpen Focus: Learning With Attention Separability and Consistency38
This paper proposes a new framework that makes class-discriminative attention a principled part of the learning process and introduces new learning objectives for attention separability and cross-layer consistency, which result in improved attention discriminability and reduced visual confusion.
- Learning View-Disentangled Human Pose Representation by Contrastive Cross-View Mutual Information Maximization37
A novel representation learning method to disentangle pose-dependent as well as view-dependent factors from 2D human poses which outperforms other competing methods by a large margin in the single-shot cross-view setting and can significantly boost the performance when reducing the amount of supervised training data.
- 37
- 37
- SnapGen-V: Generating a Five-Second Video within Five Seconds on a Mobile Device34
This work proposes a comprehensive acceleration framework to bring the power of the large-scale video diffusion model to the hands of edge users and proposes a dedicated adversarial fine-tuning algorithm for this efficient model.
- 34
- Recognition of Nonmanual Markers in American Sign Language (ASL) Using Non-Parametric Adaptive 2D-3D Face Tracking34
The proposed new framework makes it possible to detect grammatically significant nonmanual expressions from continuous signing and to differentiate successfully among linguistically significant expressions that involve subtle differences in appearance.
- Tracking Facial Features Using Mixture of Point Distribution Models34
The contribution is to apply ASM to the task of tracking shapes involving wide aspect changes and generic movements by incorporating shape priors that are learned over non-linear shape space and using them to learn the plausible shape space.
- 34
- A Learning Framework for the Automatic and Accurate Segmentation of Cardiac Tagged MRI Images34
The experimental results show that the fully automatic and accurate segmentation framework for 2D tagged cardiac MR images can achieve a highly accurate performance without any human interaction.
- Deformable-Model Based Textured Object Segmentation33
An information-theoretical approach for computing the natural scale of a “texon” based on model-interior texture, a nonparametric texture statistics comparison technique and the determination of object belongingness through belief propagation are introduced.
- Extracting tissue deformation using Gabor filter banks33
This paper presents a new approach for accurate extraction of tissue deformation imaged with tagged MR, based on banks of Gabor filters, and presents an interpolation method to recover all tag information at a finer resolution than the filter bank parameters.
- Gibbs Prior Models, Marching Cubes, and Deformable Models: A Hybrid Framework for 3D Medical Image Segmentation33
A hybrid 3D segmentation framework which combines Gibbs models, marching cubes and deformable models is proposed and it is shown that high quality segmentation can be achieved with computational efficiency.
- Autonomous Animation and Control of Four- Legged Animals33
A framework for the realistic autonomous animation and motion control of fourlegged animals using a combination of kinematics, dynamics and control theory to model walking, trotting, and simple behaviors such as target pursuit.
- Automatic Detection and Segmentation of Ground Glass Opacity Nodules32
This paper develops a classifier by boosting k-NN whose distance measure is the Euclidean distance between the nonparametric density estimates of two examples for GGO detection and presents very promising results for automatic GGO segmentation.
- 31
- From circle to 3-sphere: Head pose estimation by instance parameterization31
A coarse-to-fine pose estimation framework, where the unit circle and 3-sphere are employed to model the manifold topology on the coarse and fine layer respectively, that can uniformly factorize multiple factors in an instance parametric subspace, where novel inputs can be synthesized under a generative framework.
- Segmentation of myocardium using deformable regions and graph cuts31
A new segmentation system for (semi-) automatic delineation of epicardium and endocardium of Left Ventricle of the heart in Magnetic Resonance Images (MRI) via a coupling between deformable models and graph cuts which provides automated accurate cues for graph cuts and also good initialization scheme for de-formable model that ultimately leads to more accurate and smooth segmentation results with lower interaction costs.
- 31
- 31
- Improving Compositional Text-to-image Generation with Large Vision-Language Models30
The proposed methodology significantly improves text-image alignment in compositional image generation, particularly with respect to object number, attribute binding, spatial relationships, and aesthetic quality.
- 30
- 3D Meshless Prostate Segmentation and Registration in Image Guided Radiotherapy30
A fast segmentation-registration-segmentation framework to accurately and efficiently delineate the clinically critical objects in Cone Beam CT images obtained during radiation treatment to optimize IGRT is developed.
- Boosting and Nonparametric Based Tracking of Tagged MRI Cardiac Boundaries30
This paper extends the conventional Adaboost classifier into a posterior probability form, which can be embedded in a particle filtering-based shape tracking framework, and uses a PCA subspace shape representation to constrain the shape variation and lower the dimensionality.
- Volumetric heart modeling and analysis30
Special mechanisms in the heart provide cardiac rhythm and transmit action potentials throughout the heart muscle to cause the heart's rhythmic relaxation (diastole) and contraction (systole).
- A Novel Stochastic Combination of 3D Texture Features for Automated Segmentation of Prostatic Adenocarcinoma from High Resolution MRI30
This work presents a new methodology for fully automated segmentation of prostatic adenocarcinoma from high resolution MR by using a novel feature ensemble of 3D texture features that was found to be superior in terms of Sensitivity and Positive Predictive Value.
- Integrating Anatomy and Physiology for Behavior Modeling30
This paper describes the current Finite Element Method implementation of a simplified lung and chest cavity during normal quiet breathing and then disturbed by a simple pneumothorax, where the lung model interacts with the model of the chest cavity through applied forces.
- FR-Net: Joint Reconstruction and Segmentation in Compressed Sensing Cardiac MRI29
A novel solution to the inverse problem in medical imaging that takes as input the undersampled k-space data from Magnetic Resonance Imaging (MRI) scans and outputs both the reconstructed images and the segmented myocardium.
- 29
- 29
- Face morphing using 3D-aware appearance optimization29
This work proposes a new face morphing approach that deals explicitly with large pose and expression variations using a projection on a pre-learned 3D face subspace and poses the morphing problem as an iterative optimization with an objective.
- American Sign Language Video Anonymization to Support Online Participation of Deaf and Hard of Hearing Users28
Users’ perception of key trade-offs among these three dimensions, factors that contribute to each, and their views on transformation options enabled by this technology are revealed, to guide future designers of this technology and inform selection of applications and design features.
- Neutrophil to Lymphocyte Ratio As A Predictive Biomarker for Computed Tomography Scan Use in Mild Traumatic Brain Injury28
Investigation of the role of neutrophil-to-lymphocytes ratio (NLR) at admission for predicting the need for computed tomography (CT) in mild-TBI found that patients with positive CT-findings had significant higher NLR-levels.
- A New Framework for Sign Language Recognition based on 3D Handshape Identification and Linguistic Modeling28
A new framework is proposed that provides a new tracking method less dependent on laboratory conditions and able to deal with variations in background and skin regions, and allows for identification of 3D hand configurations that are linguistically important in American Sign Language (ASL).
- Conditional Visual Tracking in Kernel Space28
This work combines kernel PCA-based non-linear dimensionality reduction (kPCA) and Conditional Bayesian Mixture of Experts (BME) in order to learn complex multivalued predictors between observations and model hidden states for accurate, inverse, visual perception inferences.
- Joint Kernel-Based Supervised Hashing for Scalable Histopathological Image Analysis27
This paper exploits a joint kernel-based supervised hashing (JKSH) framework for fusion of complementary features and achieves 88.1% retrieval precision and 91.2% classification accuracy within 14.0 ms query time, comparing favorably with traditional methods.
- 27
- Finite Element Implementation27
The presentation examines the criteria for choosing the appropriate elements for a given problem, gives various finite element tessellations of the parametric space of a deformable model, shows examples of finite elements, and presents the finite element-based approximation to the Lagrange equations of motion.
- Incompressible Biventricular Model Construction and Heart Segmentation of 4D Tagged MRI26
This chapter presents a cardiac segmentation method for tMRI which requires no user input, suppresses image artifacts, extracts heart features using 3D grayscale morphology, and constructs a biventricular model from the data that ensures the near incompressibility of heart tissue.
- Layout-Agnostic Scene Text Image Synthesis with Diffusion Models25
The novelty of SceneTextGen lies in its integration of three key components: a character-level encoder for capturing detailed typographic properties, coupled with a character-level instance segmentation model and a word-level spotting model to address the issues of unwanted text generation and minor character inaccuracies
- 25
- 25
- 3D Motion Modeling and Reconstruction of Left Ventricle Wall in Cardiac MRI25
A novel approach for 3D motion modeling and analysis of LV wall in cardiac magnetic resonance imaging (MRI) is presented, which can visually analyze 3D LV wall motion, evaluate cardiac global function, and diagnose ventricular diseases.
- Segmentation using deformable models with affinity-based localization25
An algorithm for segmenting objects with simple closed curves, such as the heart and the lungs, that is independent of the imaging modality used is developed, which automates and significantly improves the object boundary detection results.
- 24
- Dual Projection Generative Adversarial Networks for Conditional Image Generation24
A Dual Projection GAN (P2GAN) model that learns to balance between data matching and label matching and an improved cGAN model with Auxiliary Classification that directly aligns the fake and real conditionals P (class|image) by minimizing their f-divergence.
- 24
- 24
- Medical Simulation24
This analysis demonstrates that a quasi-linear viscoelastic model fails in describing the observed evolution from the “virgin” to the preconditioned state, and is demonstrated to be unsuitable for in-vivo applications under sterile conditions during open surgery.
- 3D Cardiac Anatomy Reconstruction Using High Resolution CT Data24
An improved 3D hybrid segmentation framework is used which integrates Iibbs prior models, deformable models, and the marching cubes method to achieve a sub-pixel accuracy of the reconstruction of cardiac objects and introduces a new type of external force, which is called the scalar gradient.
- Multi-Level Shape Representation Using Global Deformations and Locally Adaptive Finite Elements24
A model-based method for the multi-level shape, pose estimation and abstraction of an object's surface from range data that ensures that during subdivision the desirable finite element mesh generation properties of conformity, non-degeneracy and smoothness are maintained.
- The state-of-the-art in cardiac MRI reconstruction: Results of the CMRxRecon challenge in MICCAI 202323
A comprehensive overview of the challenge design, a summary of the submitted results, a review of the employed methods, and an in-depth discussion that aims to inspire future advancements in cardiac MRI reconstruction models are presented.
- 23
- RED-Net: A Recurrent Encoder–Decoder Network for Video-Based Face Alignment23
This work proposes a novel method for real-time face alignment in videos based on a recurrent encoder–decoder network model that predicts 2D facial point heat maps regularized by both detection and regression loss, while uniquely exploiting recurrent learning at both spatial and temporal dimensions.
- Detection of Major ASL Sign Types in Continuous Signing For ASL Recognition23
This paper presents a multiple instance learning-based segmentation system that accurately labels 91.27% of the video frames of 500 continuous utterances from the publicly accessible NCSLGR corpus (Neidle and Vogler, 2012).
- 3D Anatomical Shape Atlas Construction Using Mesh Quality Preserved Deformable Models23
This paper proposes a new framework for 3D anatomical shape atlas construction that aims to robustly establish correspondences across different subjects and simultaneously generate high-quality surface meshes without removing shape details, and employs the Laplacian representation to encode shape details and smoothness constraints.
- Spatial and temporal pyramids for grammatical expression recognition of American sign language23
A novel framework for robust tracking and analysis of facial expression and head gestures, with an application to sign language recognition is proposed, which applies to recognition with excellent accuracy of two classes of grammatical expressions, namely wh-questions and negative expressions.
- A three-dimensional virtual environment for modeling mechanical cardiopulmonary interactions23
A real-time computer system for modeling mechanical physiological behavior in an interactive, 3-D virtual environment and a sample of results showing normal breathing and acute effects of pneumothoraces are described.
- OmniLabel: A Challenging Benchmark for Language-Based Object Detection22
A novel task definition, dataset, and evaluation metric is proposed that subsumes standard- and open-vocabulary detection as well as referring expressions for language-based object detection and judges performance via a modified average precision metric.
- PC-U Net: Learning to Jointly Reconstruct and Segment the Cardiac Walls in 3D from CT Data22
A PC-U net is proposed that jointly reconstructs the point cloud of the LV MYO wall directly from volumes of 2D CT slices and generates its segmentation masks from the predicted 3D point cloud and is beneficial for automatic cardiac image analysis tasks.
- Object-Guided Instance Segmentation for Biological Images22
A new box-based instance segmentation method that achieves state-of-the-art performances on three biological datasets: cell nuclei, plant phenotyping dataset, and neural cells by locating the object bounding boxes from their center points.
- 22
- 22
- 22
- Cascaded MRI-SPAMM for LV motion analysis during a whole cardiac cycle22
A new paradigm which incorporates multiple sets of tagged MRI data (MRI-SPAMM) acquired in a cascaded fashion in order to estimate the full 3-D motion of the left ventricle (LV) during its entire cardiac cycle is presented.
- LoR-VP: Low-Rank Visual Prompting for Efficient Vision Model Adaptation21
A novel visual prompt design is proposed, introducing Low-Rank matrix multiplication for Visual Prompting (LoR-VP), which enables shared and patch-specific information across rows and columns of image pixels.
- Dealing with Heterogeneous 3d Mr Knee Images: A Federated Few-Shot Learning Method with Dual Knowledge Distillation21
This paper proposes a federated few-shot learning method with dual knowledge distillation that allows joint training with limited annotations across clients without jeopardizing privacy and shows superior performance and less training time than other semi-supervised federated learning methods.
- Measurement of subcellular texture by optical Gabor-like filtering with a digital micromirror device21
The authors' method showed significantly rounder structures within apoptosis-defective cells lacking the proapoptotic mitochondrial effectors Bax and Bak, when compared with Bax/Bak expressing cells functional for apoptosis, consistent with reported differences in mitochondrial shape in these cells.
- Patient-specific analysis of left ventricular blood flow21
This novel approach allows for the first time a patientspecific LV blood flow simulation using exact boundary conditions as measured by MRI-SPAMM.
- A computationally efficient 3D/2D registration method based on image gradient direction probability density function20
This paper introduces a registration method which is significantly faster and less sensitive to initialization than existing methods and uses a weighted histogram of image gradient directions (WHGD) as the image feature and simplifies the computation by searching the parameter space sequentially rather than simultaneously.
- 20
- 20
- 20
- Robust segmentation of 4D cardiac MRI-tagged images via spatio-temporal propagation20
A robust method for segmenting and tracking cardiac contours and tags in 4D cardiac MRI tagged images via spatio-temporal propagation based on the Metamorphs Segmentation for robust boundary estimation and the tunable Gabor filter bank for tagging lines enhancement, removal and myocardium tracking is presented.
- Automated Model-Based Segmentation of the Left and Right Ventricles in Tagged Cardiac MRI20
An automated, model-based method to segment the left and right ventricles in 4D tagged MR through 3D epicardial and endocardial surface models to ventricle features the authors extract from the image data is described.
- LV-RV Shape Modeling Based on a Blended Parameterized Model20
This paper introduces a blended deformable model approach with parameter functions which is generic enough to deal with the different heart shapes and includes the inflow and outflow tract of the RV basal area, so that the full LV-RV shape can be estimated.
- Brain Segmentation from k-Space with End-to-End Recurrent Attention Network19
A novel learning framework that performs magnetic resonance brain image segmentation directly from k-space data and introduces a novel workflow to generate labeled training data for segmentation by exploiting imaging modality simulators and digital phantoms to address the challenge of manual labeling.
- 19
- Computer-based tracking, analysis, and visualization of linguistically significant nonmanual events in American Sign Language (ASL)19
A linguistically motivated computational approach and the results for detection and recognition of nonmanual grammatical markings are summarized and demonstrated; and work underway to enable such visualizations to be produced over large corpora and shared publicly on the Web is described.
- Shape Prior Modeling Using Sparse Representation and Online Dictionary Learning19
An online learning method that can be gracefully scaled up to model shape priors from a large number of training shapes without sacrificing run-time efficiency and is validated on lung localization in X-Ray and cardiac segmentation in MRI time series.
- Tracking the Swimming Motions of C. elegans Worms with Applications in Aging Studies19
A novel training-based method that automatically tracks and segments multiple swimming worms, in challenging imaging conditions, allowing a quantitative comparison of worm populations over their lifetime.
- A new spatio-temporal MRF framework for video-based object segmentation19
A general framework for videobased object segmentation using a new spatio-temporal Markov Random Field (MRF) model that includes the inclusion of shape priors as a top-down highlevel object constraints as guides for the bottom-up low-level image cues leads to improved segmentation.
- 19
- Outlier rejection in high-dimensional deformable models19
This paper presents three simple, yet effective ways to find the outliers in deformable model tracking, and validate and compare these approaches in an 11-parameter deformable face tracking application against ground truth data.
- 18
- 18
- 18
- Greedy AutoAugment18
This paper proposes Greedy AutoAugment as a highly efficient search algorithm to find the best augmentation policies and uses a greedy approach to reduce the exponential growth of the number of possible trials to linear growth.
- Towards Image-to-Video Translation: A Structure-Aware Approach via Multi-stage Generative Adversarial Networks18
This work proposes a two-stage generative framework where videos are forecast from the structural sequence and then refined by temporal signals, which avoids learning motion-irrelevant details in the refining stage.
- A Method for Recognition of Grammatically Significant Head Movements and Facial Expressions, Developed Through Use of a Linguistically Annotated Video Corpus 118
This corpus of linguistically annotated video data collected from native ASL signers is accessible for use by the linguistics and computer science research communities and for educational purposes through a new Web-based Data Access Interface (DAI) that is developing to facilitate viewing, searching, and downloading relevant subsets of the data.
- Highly sensitive size discrimination of sub-micron objects using optical Fourier processing based on two-dimensional Gabor filters18
The optical Gabor-like filtering implemented with a digital micromirror device to achieve nanoscale sensitivity to changes in the size of finite and periodic objects imaged at low resolution may have applications in high throughput optical analysis of subcellular morphology to study organelle function in living cells.
- Identifying Regional Cardiac Abnormalities from Myocardial Strains Using Spatio-temporal Tensor Analysis18
A novel tensor-based classification framework is developed that better conserves the spatio-temporal structure of the myocardial deformation pattern than conventional vector-based algorithms and keeps more of the information of the original feature space, so that abnormal tensors in the subspace can be back-projected to reveal the regional cardiac abnormality in a more physically meaningful way.
- Generating Enhanced Negatives for Training Language-Based Object Detectors17
This work uses large-Language-models to generate negative text descriptions, and text-to-image diffusion models to also generate corresponding negative images, and proposes to leverage the vast knowledge built into modern generative models to automatically build negatives that are more relevant to the original data.
- Adaptive low rank and sparse decomposition of video using compressive sensing17
A new method that performs video reconstruction by low rank and sparse decomposition adaptively by using a background model in which the background is learned adaptively as the compressive measurements are processed.
- Hybrid System Models of Navigation Strategies for Games and Animations17
This work concretely applies hybrid system theory and tools to model navigation strategies for virtual characters and suggests that hybrid systems may be a natural framework for modeling aspects of intelligent virtual actors.
- Improved Training Technique for Latent Consistency Models16
This work replaces Pseudo-Huber losses with Cauchy losses and introduces the adaptive scaling-$c$ scheduler to manage the robust training process and adopt Non-scaling LayerNorm in the architecture to better capture the statistics of the features and reduce outlier impact.
- DeFormer: Integrating Transformers with Deformable Models for 3D Shape Abstraction from a Single Image16
A novel bi-channel Transformer architecture, integrated with parameterized deformable models, termed DeFormer, to simultaneously estimate the global and local deformations of primitives, and achieves better reconstruction accuracy over the state-of-the-art.
- AE-StyleGAN: Improved Training of Style-Based Auto-Encoders16
This paper describes a new methodology to train a style-based autoencoder where the encoder and generator are optimized end-to-end and shows that the proposed model consistently outperforms baselines in terms of image inversion and generation quality.
- Linguistically-driven Framework for Computationally Efficient and Scalable Sign Recognition16
A new general framework for sign recognition from monocular video using limited quantities of annotated data that exploits state-of-the art learning methods while also incorporating features based on what the authors know about the linguistic composition of lexical signs.
- Reconstruction for Feature Disentanglement in Pose-invariant Face Recognition16
This paper presents a method for learning a feature representation that is invariant to pose, without requiring extensive pose coverage in training data, and proposes a Siamese network to explicitly disentangle identity and pose.
- Mode Estimation for High Dimensional Discrete Tree Graphical Models.16
Though the mode finding problem is generally intractable in high dimensions, this paper unveils that, if the distribution can be approximated well by a tree graphical model, mode characterization is significantly easier.
- A Framework for the Recognition of Nonmanual Markers in Segmented Sequences of American Sign Language16
This work proposes a novel framework for robust tracking and analysis of non-manual behaviours, with an application to sign language recognition, and applies it to segmented recognition of five classes of grammatical constructions conveyed through facial expressions and head gestures.
- 16
- 15
- 15
- Accurate segmentation of brain images into 34 structures combining a non-stationary adaptive statistical atlas and a multi-atlas with applications to Alzheimer'S disease15
An extended adaptive statistical atlas method (EASA) is constructed to use a non-stationary relaxation factor rather than a global one, which permits finer control over adaptivity allowing 34 structures to be simultaneously segmented rather than just 4 as in [13].
- Dynamic soft encoded patterns for facial event analysis15
This paper develops similarity features to describe complicated variations of facial appearance, which take similarities between a haar-like feature in a given image and the corresponding ones in reference images as feature vector, and maps them into dynamic soft encoded pattern (DSEP) for facial event analysis.
- Learning-based dynamic coupling of discrete and continuous trackers15
A data-driven dynamic coupling between discrete and continuous methods for tracking objects of high dofs, which overcomes the limitations of previous techniques and applies the framework to the problem of 3D hand tracking from image sequences and the discrimination between fingerspelling and continuous signs in American Sign Language.
- Inferring 2D Object Structure from the Deformation of Apparent Contours15
A new algorithm for force assignment that assigns forces from the data to multiple models that allows partial overlap between the parts' models and the determination of their joint location is presented.
- More Than Just Attention: Improving Cross-Modal Attentions with Contrastive Constraints for Image-Text Matching14
This work proposes two novel training strategies, namely Contrastive Content Resourcing (CCR) and Contrastive content Swapping (CCS) constraints, to address limitations in cross-modal attention models and introduces three metrics, including Attention Precision, Recall, and F1-Score, to quantitatively measure the quality of learned attention models.
- Automatic tooth segmentation from 3D dental model using deep learning: a quantitative analysis of what can be learnt from a single 3D dental model14
The first to quantitatively evaluate and demonstrate the representation learning capability of Deep Learning methods from a single 3D intraoral scan can enable building self-supervision methods for tooth segmentation under extreme data limitation scenario by leveraging the available data to the fullest possible extent.
- 14
- Sparse Deformable Models with Application to Cardiac Motion Analysis14
This work introduces a new family of deformable models that are inspired from compressed sensing, a technique for efficiently reconstructing a signal based on its sparseness in some domain, and applies it to the analysis of cardiac motion using tagged magnetic resonance imaging (tMRI), where the automated tagging line tracking results are very noisy due to the poor image quality.
- 14
- A 3D global‐to‐local deformable mesh model based registration and anatomy‐constrained segmentation method for image guided prostate radiotherapy14
A novel and fast segmentation and deformable registration method was developed to capture the transformation between the planning and treatment images for external beam radiotherapy of prostate cancers that increases the computational efficiency and may provide foundation to achieve real time adaptive radiotherapy.
- Human Motion: Understanding, Modelling, Capture, and Animation (Computational Imaging and Vision)14
Researchers from all three of these areas have contributed to this book to promote the establishment of human motion research as a multi-facetted discipline and to improve the exchange of ideas and concepts between these three areas.
- 14
- Scan-Conversion Algorithm for Ridge Point Detection on Tubular Objects14
This paper presents a gradient-based method for ridge point extraction on tubular objects that is efficient and allows for completely automatic extraction of points along the centerline of a tubular object in its elongated direction.
- Efficient Dynamic Constraints for Animating Articulated Figures14
An efficient dynamics-based computer animation system for simulating and controlling the motion of articulated figures is presented and a non-trivial extension of Featherstone's O(n) recursive forward dynamics algorithm is derived which allows enforcing one or more constraints on the animated figures.
- Toward Personalized Modeling: Incremental and Ensemble Alignment for Sequential Faces in the Wild13
This work samples multiple initial shapes to achieve image congealing within one frame, which enables us to incrementally conduct ensemble alignment by group-sparse regularized rank minimization and leverages a very efficient fitting evaluation network to pick out well-aligned faces for robust incremental learning.
- 13
- 3D/2D image registration using weighted histogram of gradient directions13
A novel feature-based registration method using the weighted histogram of gradient directions of images which can tolerate up to ±90°rotation offset from the global optimal solution, which minimizes the need for human interaction to initialize the algorithm.
- 13
- 3D Segmentation of Rodent Brain Structures Using Hierarchical Shape Priors and Deformable Models13
This method combines deformable models and hierarchical shape priors within one framework to segment multiple rodent brain structures simultaneously and this hierarchical prior improves the segmentation performance.
- Lesion-Specific Coronary Artery Calcium Quantification for Predicting Cardiac Event with Multiple Instance Support Vector Machines13
A lesion-specific CAC quantification framework to improve CAC's near-term predictive value in intermediate to high-risk populations with a novel multiple instance support vector machines (MISVM) approach is proposed.
- 13
- Learning to Localize Actions in Instructional Videos with LLM-Based Multi-pathway Text-Video Alignment12
This work first applies an LLM to filter out task-irrelevant information and summarize task-related procedure steps (LLM-steps) from narrations to generate reliable pseudo step-video matching, and proposes the Multi-Pathway Text-Video Alignment (MPTVA) strategy, which surpasses state-of-the-art methods in three downstream tasks.
- 12
- DIAGNOSIS: Detecting Unauthorized Data Usages in Text-to-image Diffusion Models12
This paper alters the protected image dataset by adding unique contents on the images such as stealthy image wrapping functions that are imperceptible to human vision but can be captured and memorized by diffusion models to detect unauthorized data usages.
- 12
- Heuristic Search for Homology Localization Problem and Its Application in Cardiac Trabeculae Reconstruction12
This work forms the problem of accurate extraction of cardiac trabeculae as a heuristic search problem, and proposes novel heuristic functions based on advanced topological techniques that improve the computation in both time and memory.
- 12
- Dynamic MRI Reconstruction with Motion-Guided Network12
A novel dynamic MRI reconstruction approach called MODRN is proposed that unitizes deep neural networks with motion information to improve reconstruction quality and demonstrates the effectiveness of the proposed approach compared to other state-of-the-art approaches.
- The Importance of 3D Motion Trajectories for Computer-based Sign Recognition12
This framework demonstrates the discriminative power and importance of 3D hand motion trajectories for sign recognition, given known handshapes, given known handshapes.
- 12
- Simplified Labeling Process for Medical Image Segmentation12
A robust logistic regression algorithm is proposed to handle label outliers such that doctors do not need to waste time on precisely labeling images for training set, which validates the benefits of the proposed algorithms.
- Spectrum-Aware Parameter Efficient Fine-Tuning for Diffusion Models11
Spectral Orthogonal Decomposition Adaptation (SODA), which balances computational efficiency and representation capacity, is introduced, offering a spectrum-aware alternative to existing fine-tuning methods.
- 11
- Neural Deformable Models for 3D Bi-Ventricular Heart Shape Reconstruction and Modeling from 2D Sparse Cardiac Magnetic Resonance Imaging11
A novel neural deformable model (NDM) targeting at the reconstruction and modeling of 3D bi-ventricular shape of the heart from 2D sparse cardiac magnetic resonance (CMR) imaging data that can learn to densify a sparse cardiac point cloud and generate high-quality triangular meshes automatically.
- 11
- GNM: GridCell navigational model11
A novel computational model for patterns of grid cells is proposed and combined with a mechanism to tune the weights of cells, which is used to create a decision-making process for robot navigation.
- Cartoonish sketch-based face editing in videos using identity deformation transfer11
A two-stage real-time 3D face model fitting schema is introduced to recover facial identity and expressions from the video and a novel identity deformation algorithm is proposed to transfer deformations from 2D space to 3D facial identity directly, while preserving the facial expressions.
- Cervigram image segmentation based on reconstructive sparse representations11
In the experiments, the method to automatically segment the biomarker AcetoWhite regions in an archive of 60,000 images of the uterine cervix showed lower space and time complexity and higher sensitivity.
- Active Volume Models with Probabilistic Object Boundary Prediction Module11
The AVM thus has the efficiency of Snakes while having adaptive region-based constraints and is a natural extension of the Snakes model so that region information becomes part of the external forces.
- Shape and Nonrigid Motion Estimation11
In their previous presentation of the shape and nonrigid motion estimation algorithm, the authors did not incorporate formally in the estimation process the noise in the data, but in this presentation they do so formally.
- Toward Modality- and Sampling-Universal Learning Strategies for Accelerating Cardiovascular Imaging: Summary of the CMRxRecon2024 Challenge10
Analysis of the best-performing solutions revealed that prompt-based adaptation and enhanced physics-driven consistency enabled strong cross-scenario performance and establish principles for generalizable reconstruction models and advance clinically translatable AI in cardiovascular imaging.
- 10
- Deformable models, sparsity and learning-based segmentation for cardiac MRI based analytics10
This chapter presents a new deep learning method to handle segmentation of multiple structures in cine CMR data, and proposes a modified deep layer aggregation architecture with channel attention and refinement residual blocks to better fuse appearance information across layers during training and achieve improved results through multiscale analysis of image appearance.
- Effective 3D humerus and scapula extraction using low-contrast and high-shape-variability MR data10
This paper presents a joint segmentation for the humerus and scapula bones on a small dataset with low-contrast and high-shape-variability 3D MR images and designs a self-reinforced learning strategy to increase performance.
- Temporal Shape Analysis via the Spectral Signature10
This paper proposes a simple encoding motivated by the observation that small shape deformations lead to minor refinements in the spectral signature composed of the eigenvalues of the Laplace operator.
- 10
- 10
- Fast Motion Tracking of Tagged MRI Using Angle-Preserving Meshless Registration10
A new fast motion tracking method based on a meshless kernel that preserves tangential angles of tags at tag intersections and generates a dense displacement map that minimizes the residual of sparse motion at intersections is presented.
- 10
- 9
- Instance Segmentation of Neural Cells9
This work proposes a novel deep multi-task learning model to jointly detect and segment neural cells instance-wise, built upon SSD, with ResNet101 as the backbone to achieve both high detection accuracy and fast speed.
- 9
- Fast generation of digitally reconstructed radiograph through an efficient preprocessing of ray attenuation values9
A new preprocessing procedure and data structure for the calculation of the ray attenuation values is proposed, which supports all possible volume positions with practically small memory requirements in addition to reducing the complexity of the problem from O(n3) to O( n2).
- Comparison of windowing effects on elastography images: Simulation, phantom and in vivo studies9
It is hypothesize that the use of smooth windows may improve the elastographic signal-to-noise ratio (SNRe) and contrast- to-noISE ratio (CNRe) of strain images and "Salt-and-pepper" noise from false-peak errors has also been reduced.
- Detection of Mammographic Masses by Content-Based Image Retrieval9
This work proposes a CAD approach that integrates image retrieval with learning-based mass detection, and achieves a promising ROC area index A z = 0.91, and outperforms two traditional classifier-based CAD methods.
- 9
- Integrating Shape and Texture in 3D Deformable Models: From Metamorphs to Active Volume Models9
This chapter reviews a 2D deformable model – Metamorphs, which integrates region texture constraints so as to achieve more robust segmentation, and Active Volume Model (AVM), a similar and improved approach for 3D segmentation.
- Tag Separation in Cardiac Tagged MRI9
This paper introduces a tag separation method for better cardiac boundary segmentation and tag tracking, and demonstrates the superior performance of the proposed approach through extensive experiments on large sets of cardiac tagged MR images.
- 9
- Generation, estimation and tracking of faces9
This thesis describes techniques for the construction of face models for both computer graphics and computer vision applications and derives and solves a dynamic system which accounts for edges in the image and incorporates optical flow as a motion constraint on the model.
- DiffSLVA: Harnessing Diffusion Models for Sign Language Video Anonymization8
DiffSLVA is introduced, a novel methodology that utilizes pre-trained large-scale diffusion models for zero-shot text-guided sign language video anonymization that could be used for real-world applications, which would offer significant benefits to the Deaf and Hard-of-Hearing communities.
- 8
- Understanding ASL Learners’ Preferences for a Sign Language Recording and Automatic Feedback System to Support Self-Study8
Ten participants in this study were asked to record a few sentences in ASL using software designed, and they provided manually curated feedback on one sentence in a manner that simulates the output of a future automatic feedback system.
- Sparse Data–Driven Learning for Effective and Efficient Biomedical Image Segmentation8
A selection of techniques that incorporate sparsity, including strategies based on dictionary learning and deep learning, that are aimed at medical image segmentation and related quantification are presented.
- 8
- 8
- 8
- Parallel Sparse Subspace Clustering via Joint Sample and Parameter Blockwise Partition8
A memory-efficient parallel framework for computing SSC via an alternating direction method of multiplier (ADMM) algorithm that is substantially more efficient than its matrix counterpart used by SSC, without sacrificing performance in applications.
- 8
- 8
- Similarity Features for Facial Event Analysis8
This paper proposes similarity features to describe the facial appearance for video-based facial event analysis, inspired by the kernel features, and maps the similarity features into dynamic binary patterns to handle the issue of time resolution.
- Variational Bayesian Sequence-to-Sequence Networks for Memory-Efficient Sign Language Translation7
This work introduces a paradigm of designing recurrent deep networks whereby the output of the recurrent layer is derived from appropriate arguments from nonparametric statistics, and proposes a novel variational Bayesian sequence-to-sequence network architecture.
- Track Facial Points in Unconstrained Videos7
This paper proposes to exploit incremental learning for person-specific alignment in wild conditions using part-based representation and cascade regression for robust and efficient alignment on each frame to achieve personalized modeling on the fly.
- Multi-layer stencil creation from images7
This paper proposes a system which separates the artistic design stage from the complex and tedious task of stencil creation, formulated as a constrained energy optimization problem and solved efficiently.
- 7
- 7
- Accurate Whole-Brain Segmentation for Alzheimer’s Disease Combining an Adaptive Statistical Atlas and Multi-atlas7
A framework combining the improved adaptive statistical atlas with a multi-atlas method which achieves simultaneous accurate segmentation of the cortex, ventricles, and sub-cortical structures in severely diseased brains, a feat not attained in [18].
- 7
- 7
- 7
- Dynamic Data Driven Coupling of Continuous and Discrete Methods for 3D Tracking7
A new framework for robust 3D tracking is presented, using a dynamic data driven coupling of continuous and discrete methods to overcome their limitations and simultaneously reduce the discrete method's complexity and prevent possible multiple solutions in shape estimation.
- A Framework for Reasoning About Animation Systems7
This paper begins by directly applying hybrid systems theory to animation, using a general-purpose hybrid system specification tool to generate multi-agent animations; this application illustrates that hybrid system models can provide systematic, modular ways to incorporate low-level behavior into a design for higher-level behavioral modeling.
- Image-based ventricular blood flow analysis7
A novel and non-invasive method for the quantitative analysis of blood flow in the left ventricle of specific human patients using information derived from medical images and integrating region-based and physics-based modeling techniques is described.
- SODA: Spectral Orthogonal Decomposition Adaptation for Diffusion Models6
Spectral Orthogonal Decomposition Adaptation (SODA), which balances computational efficiency and representation capacity, is introduced, offering a spectrum-aware alternative to existing fine-tuning methods.
- A Critical Analysis of the Limitation of Deep Learning based 3D Dental Mesh Segmentation Methods in Segmenting Partial Scans6
It is discovered that the high performing tooth segmentation methods under-perform when segmenting partial intraoral scans and the analysis and comparison presented in this work would help in understanding the severity of the problem and allow to develop robust Tooth segmentation technique without strong assumption of full jaw model.
- 6
- 6
- 6
- Visual tracking with reliable memories6
A novel visual tracking framework that intelligently discovers reliable patterns from a wide range of video to resist drift error for long-term tracking tasks is proposed and a clustering method with temporal constraints to explore and memorize consistent patterns from previous frames, named as "reliable memories".
- 6
- 6
- A Unified Framework of Subspace and Distance Metric Learning for Face Recognition6
This paper proposes a unified scheme of subspace and distance metric learning under the Bayesian framework for face recognition, and formulate it by minimizing the Bhattachyrra coefficient between two distributions.
- 6
- Characterization of Time-Varying Properties and Regional Strains in Myocardial Ischemia6
This review outlines the steps to deduct myocardial mechanical performance from medical imaging modalities and modeling and uses time-varying elastance model instead of elastic solid continuum model to provide a more accurate description of regional myocardium mechanics.
- Feel the fabric via the phantom6
From the user evaluation, it is concluded that the surface roughness and friction of the four different virtual fabrics can be perceived haptically and ordered by the subjects.
- 6
- Constructive Assimilation: Boosting Contrastive Learning Performance through View Generation Strategies5
This work proposes a view generation method and a simple, effective assimilation method that together improve the state-of-the-art by up to ~3.6% on three different datasets.
- Modality Bank: Learn multi-modality images across data centers without sharing medical data5
This work proposes privacy secured decentralized multi-modality adaptive learning architecture named ModalityBank, which could learn a set of effective domain-specific modulation parameters plugged into a common domain-agnostic network and could also complete the missing modalities across all data centers.
- 5
- 5
- Detecting Mammographic Masses via Image Retrieval and Discriminative Learning5
This work introduces an automatic CAD approach that integrates image retrieval and discriminative learning and achieves high mass detection accuracy and retrieval precision, comparing favorably with traditional methods.
- 3D Face Tracking and Multi-Scale, Spatio-temporal Analysis of Linguistically Significant Facial Expressions and Head Positions in ASL5
New methods for the recognition of nonmanual grammatical markers in American Sign Language (ASL) based on a 2-level learning framework to combine low- and high-level features of differing spatio-temporal scales are presented.
- 5
- 5
- 5
- 5
- 5
- Adaptive Metamorphs Model for 3D Medical Image Segmentation5
An adaptive model-based segmentation framework, in which edge and region information are integrated and used adaptively while a solid model deforms toward the object boundary, is introduced, which is more efficient and more robust to image noise, artifacts and intensity inhomogeneity.
- Registration of Lung Tissue Between Fluoroscope and CT Images: Determination of Beam Gating Parameters in Radiotherapy5
A new registration method on deformable soft tissue between fluoroscopic images and DRR (Digitally Reconstructed Radiograph) images from planning CT images using active shape models is proposed and preliminary results show that the method is very robust for the registration of deformablesoft tissue.
- A New Combinatorial Approach to Supervised Learning: Application to Gait Recognition5
A new approach that identifies similar observations across objects and use only informative data for classification is presented, which results in improved recognition rates in most cases compared with other previously reported methods.
- Incremental Model-Based Estimation Using Geometric Consistency Constraints5
A physics-based deformable model framework for the incremental object shape estimation and tracking in image sequences is presented and the forward transfer under the action of the Euclidean group and Jacobian matrices for the underlying line feature mapping is derived.
- Dynamic Simulation for Zero-Gravity Activities5
This work is developing a fast recursive dynamics algorithm for dynamically simulating articulated 3D human models, which comprises kinematic chains -serial, closed-loop, and tree-structure -as well as the inertial properties of the segments.
- Continuous Spatio-Temporal Memory Networks for 4D Cardiac Cine MRI Segmentation4
Results of extensive experiments across multiple cMR datasets show that the proposed CSTM network can improve the 4D cMR segmentation performance, especially for the hard-to-segment regions.
- Development and evaluation of a deep learning model for oral soft and hard tissue diagnosis using synthetic image data generated by StyleGAN4
A framework that utilizes pseudo-labeling to incorporate the StyleGAN2-synthesized 2D intraoral images into a tooth type classification model is created and demonstrated that the StyleGAN2 synthesized images can effectively augment the training set and improve the performance of the tooth type classification model.
- 4
- 4
- 4
- 4
- 4
- 3D LV Probabilistic Segmentation in Cardiac MRI Using Generative Adversarial Network4
This paper makes the first attempt to use probabilistic segmentation of LV myocardium for 3D heart wall reconstruction from 2D cardiac cine MRI data, to the best of the knowledge.
- Sketch-based Face Editing in Video Using Identity Deformation Transfer.4
A two-stage real-time 3D face model fitting schema is introduced to recover facial identity and expressions from the video and a novel identity deformation algorithm is proposed to transfer deformations from 2D space to 3D facial identity directly, while preserving the facial expressions.
- 4
- Localization of Multi-pose and Occluded Facial Features via Sparse Shape Representation4
This paper proposes a unified framework to robustly locate multi-pose and occluded facial features by using a sparse linear combination to approximate the observed shape, and hence alleviate the multi-posing problem.
- Construction of Left Ventricle 3D Shape Atlas from Cardiac MRI4
This paper derives a framework that creates a 3D object mesh from a 2D stack of contours, based on geometry processing algorithms and a semi-constrained deformation method, that generates subject-specific 3D models, but with global point correspondences.
- Active Contours with Level-Set for Extracting Feature Curves from Triangular Meshes4
A novel algorithm that extracts feature curves from triangular mesh domains by defining the energy functional of the active contour over the domain of triangular mesh and derive a level-set evolution equation that finds feature regions.
- A Hybrid Framework for Image Segmentation Using Probabilistic Integration of Heterogeneous Constraints4
Experimental results show that the new hybrid method outperforms methods that use homogeneous constraints and the fully probabilistic nature of the model allows us to study the utility of different inference methods and schedules.
- 4
- Curved path human locomotion on uneven terrain4
The system generates gait for different sized figures in real-time, and allows for curved path locomotion on uneven terrain, with flexibility for stylistic variation, and the solution to the inverse motion interpolation problem is described.
- 3
- DICE: Discrete Inversion Enabling Controllable Editing for Multinomial Diffusion and Masked Generative Models3
DICE (Discrete Inversion for Controllable Editing), the first approach to enable precise inversion for discrete diffusion models, including multinomial diffusion and masked generative models, is introduced.
- 3
- 3
- 3
- 3
- Unifying Subspace and Distance Metric Learning with Bhattacharyya Coefficient for Image Classification3
Inspired by the Bayesian classification error estimation, the objective function is formulated by minimizing the Bhattachyrra coefficient between two distributions and extended with the kernel trick to learn nonlinear distance metric.
- 3
- 3
- 3
- 3
- Dynamic Tracking of Facial Expressions Using Adaptive, Overlapping Subspaces3
A DDDAS to track 2D shapes across large pose variations by learning non-linear shape manifold as overlapping, piecewise linear subspaces and demonstrates the results on tracking facial features and provides several empirical results to validate the approach.
- 3
- Large Sign Language Models: Toward 3D American Sign Language Translation2
Large Sign Language Models (LSLM) is presented, a novel framework for translating 3D American Sign Language (ASL) by leveraging Large Language Models (LLMs) as the backbone, which can benefit hearing-impaired individuals’ virtual communication.
- Resolving Inconsistent Semantics in Multi-Dataset Image Segmentation2
This work introduces a simple yet effective multi-dataset training approach by integrating language-based embeddings of class names and label space-specific query embeddings of class names and label space-specific query embeddings that maintains high performance regardless of the underlying inconsistencies between training datasets.
- 2
- 2
- 2
- Measure Anatomical Thickness from Cardiac MRI with Deep Neural Networks2
Novel methods for dense thickness estimation are proposed, including a fast solver that estimates thickness from binary annular shapes and an end-to-end network that Estimates thickness directly from raw cardiac images.
- 2
- eCurves: A Temporal Shape Encoding2
A framework for temporal shape analysis to capture the shape and changes of anatomical structures from three-dimensional+t(ime) medical scans that only assumes pose invariance within a time series but does not assume point-to-point correspondence across time series or a (statistical or physical) model.
- Cardiac Trabeculae Segmentation: an Application of Computational Topology (Multimedia Contribution)2
A topological approach is developed, which employs advanced computational topology methods and achieves high quality segmentation results in cardiac trabeculae segmentation.
- 2
- Calibrationless Parallel Dynamic MRI with Joint Temporal Sparsity2
A novel calibrationless method for parallel dynamic magnetic resonance imaging (MRI) reconstruction is proposed, which overcomes the limitations posed by traditional MRI reconstruction methods that require accurate coil calibration.
- 2
- 2
- High resolution cardiac shape registration using Ricci flow2
This paper proposes a novel method to match two models with non-trivial topology using discrete hyperbolic Ricci flow and shows the performance on the registration of high resolution left ventricle models.
- 2
- 2
- 2
- 2
- 2
- Image and video classification2
New approaches for various image and video classification problems, including static image categorization, dynamic gait recognition, and unsupervised motion layer recovery are described, which explicitly identify the uninformative instances and remove them from the further classification stage.
- Recovering 3D tumor locations from 2D bioluminescence images and registration with CT images2
A novel and efficient algorithm for reconstructing the 3D locations of tumor sites from a set of 2D bioluminescence images which are taken by a same camera but after continually rotating the object by a small angle is introduced.
- Automated volumetric model construction and dynamic segmentation of the heart ventricles in tagged mri2
This work has developed a fully automated system which constructs a biventricular patient-specific model of the heart and dynamically segments the ventricles throughout systole and replaces the 4 to 17 hours of interactive segmentation with just 1.75 hours of CPU time.
- Statistical cue estimation for model-based shape and motion tracking2
A method to estimate the probability distributions of the observations in a deformable model tracking system and the use directed acyclic graphs as powerful data structure to represent and implement deformable models is introduced.
- 2
- Modeling fluid phenomena for computer graphics and special effects2
A new animation technique for modeling the turbulent rotational motion that occurs when liquids and gases interact with solid objects and the surrounding medium is described, that can handle effects such as swirling steam, rolling or billowing smoke, gusting wind, ocean wave motion, fountains, and simple explosions.
- 1
- 1
- 1
- Learning Volumetric Neural Deformable Models to Recover 3D Regional Heart Wall Motion from Multi-Planar Tagged MRI1
A novel class of volumetric neural deformable models ($\upsilon$NDMs) are introduced, which represent heart wall geometry and motion through a set of low-dimensional global deformation parameter functions and a diffeomorphic point flow regularized local deformation field.
- New Capability to Look Up an ASL Sign from a Video Example1
A new system is described, publicly shared on the Web, to enable lookup of a video of an ASL sign, enabling the user to directly look up a sign in the video being annotated, and to directly enter into the annotation the gloss and features of that sign, greatly increasing the efficiency and consistency of linguistic annotations of ASL video data.
- 3D/2D image registration method for joint motion analysis using low-quality images from mini C-arm machines1
A novel method for kinematic analysis of joint movement using the mini C-arm, where the bone of interest is recovered and isolated from the rest of the image using a non-rigid registration of an atlas to each frame, making this application available to lower-budget environments.
- Automatic stereoscopic video generation based on virtual view synthesis1
An automatic and robust framework to synthesize stereoscopic videos from casual 2D monocular videos and a Bayesian-based View Synthesis approach is proposed to render high-quality new virtual views for stereoscopic video to deal with noisy 3D geometry information.
- Auto-encoding of discriminating morphometry from cardiac MRI1
Compared with volumetric measurements and a deformation-based score, this encoding is much more accurate in capturing morphome-tric patterns distinguishing healthy subjects from those with Tetralogy of Fallot, diastolic dysfunction, and hypertrophic cardiomyopathy.
- 1
- Segmentation and Blood Flow Simulations of Patient-Specific Heart Data1
This chapter presents a fully automatic and accurate segmentation framework for 2D cardiac tagged MR images, a semiautomatic method for 3D segmentation from CT data, and the results of blood flow simulation using these highly detailed models.
- Invited Talk: Coupling Deformable Models and Learning Methods for Nonverbal Behavior Analysis: Applications to Deception, Multi-cultural Studies and ASL1
A novel system for real-time facial and gesture tracking and action recognition by using deformable statistical models that encode facial shape variation and local texture distribution, which robustly tracks 79 facial landmarks, which correspond to facial components such as the eyes, eyebrows, nose, and mouth.
- Content quality based image retrieval with multiple instance boost ranking1
This paper proposes a new idea for image retrieval, in which image retrieval is regarded as a ranking issue by evaluating image content quality by developing Mean, Max and TopK ranking models developed with Boost learning.
- 1
- 3D-3D Tubular Organ Registration and Bifurcation Detection from CT Images1
The bifurcation points (or the branching points) have been chosen to build the validation protocol of the registration methods and many researchers have developed various methods for registration of tubular organs from medical images, all developed for applications with 2D-3D registration.
- 1
- 1
- 1
- Face Mis-alignment Analysis by Multiple-Instance Subspace1
Experimental results show that the proposed learning algorithm can improve the robustness of current methods with poorly aligned training and testing data.
- 1
- Subtle Facial Expression Synthesis using Motion Manifold Embedding and Nonlinear Decomposable Generative Models1
A new framework to model facial motions of different people with multiple expression types from high resolution facial expression tracking data is presented and can synthesize high resolution facial motions based on tracking of facial motions in video sequences by estimation of the model parameters.
- Efficient Learning by Combining Confidence-Rated Classifiers to Incorporate1
A new dynamic learning framework that requires a small amount of labeled data in the beginning, then incrementally discovers informative unlabeled data to be hand-labeled and incorporates them into the training set to improve learning performance.
- In-vivo motion analysis of bi-ventricular hearts from tagged MR images1
While significant shape and motion variations were apparent in normal subjects, the quantitative analysis show typical patterns and the principal strain analysis describes the directions and magnitudes of maximum shortening, and maximum thickening.
- 1
- Medical Simulation: International Symposium, ISMS 2004, Cambridge, MA, USA, June 17-18, 2004, Proceedings1
This work focuses on the development of a framework for Biomechanical Simulation of Cranio-Maxillofacial Surgery Interventions, and the application of Embedded and Tubular Structure to Tissue Identification for the Computation of Patient-Specific Neurosurgical Simulation Models.
- Matching and recognition of shapes using chord-based point density graphs1
A novel approach to matching and recognition of 2D shapes, which uses a chord-based shape representation and matches shapes by matching correspondences be-tween chords, and employs a hierarchical approach to extend the method to include Affine transformations.
- 1
- Motion-Based Part Segmentation and Tracking1
This chapter extends the previously developed PBM framework to allow for simultaneous part segmentation, shape and motion estimation of complex multipart objects, including humans, to address a variety of problems in computer vision and computer graphics.
- Kinematics and Dynamics1
A force-based estimation technique for shape and nonrigid motion estimation in computer vision applications and a systematic procedure for converting geometrically defined parameters into physical degrees of freedom are presented.
- 1
- –
- –
- –
- –
- –
- –
- The future of automated capture of social kinesic signals for psychiatric purposes–
This article considers how computer vision can be enlisted for biomedical applications, specifically the measurement, data analytics and treatment of psychiatric disorders, and reports on various tools for analyzing social kinesic signals for emotional and physiological states.
- –
- –
- Automated Pulmonary Fibrosis Segmentation Using a 3D Multi-Scale Convolutional Encoder-Decoder Approach in Thoracic CT for the Rhesus Macaque with Radiation-Induced Lung Damage–
The qualitative and quantitative comparisons show that the proposed 3D multi-scale convolutional encoder-decoder segmentation method can achieve better segmentation accuracy with less variance in testing data.
- Video Story Question Answering with Character-Centric Scene Parsing and Question-Aware Temporal Attention–
Video Story Question Answering (VSQA) proves to be an effective way for benchmarking the comprehension ability of a model and relations and actions acquired by scene parsing are indispensable in the comprehension of video stories.
- –
- –
- –
- –
- –
- –
- Conditional Models for Three-Dimensional Human Pose Estimation.–
This thesis works with “bottom-up methods” for human pose estimation, that use low level image features to directly predict 3d pose, and incorporates unlabeled data in a semi-supervised learning framework to constrain and improve the training of discriminative models.
- –
- - Lung Nodule and Tumor Detection and Segmentation–
Stents are tubular structures used to return normal breathing function to a patient by pushing the narrowed regions of the trachea and reduce the surgical risk to the patient, especially when compared with tracheal resection.
- Registered 3D Tagged MRI and Ultrasound Myocardial Elastography: Quantitative Strain Comparison–
This chapter developed a semiautomatic intensity and gradient-based registration framework that rigidly registers the 3D tagged MRIs with the 2D ultrasound images and concludes that qualitatively tMRI is more robust and accurate than UME.
- –
- –
- –
- Nonlinear Dynamic Shape and Appearance Models for Facial Motion Tracking–
The proposed model provides a generative model for Bayesian tracking of facial motions using particle filtering with simultaneous estimation of the expression type and estimates the geometric transformation and the global deformation using the generative models.
- Shape and appearance information integration in medical image analysis and computer vision–
Novel and efficient algorithms are developed that coherently integrate shape and appearance information to augment traditional shape-only deformable models for robust segmentation, to better perform shape alignment and registration in arbitrary dimension, and to build a new framework for learning statisticalshape and appearance models that requires significantly less human effort.
- False-positive elimination for computer-aided detection of pulmonary micronodules–
A 3D filter-based method for pulmonary micronodule detection from high-resolution 3D chest CT images and a false-positive elimination method based on a deformable model, both of which focus on the automatic detection of small nodules less than 5mm in diameter.
- Modeling, simulation and analysis of the heart from four-dimensional cardiac tagged-mr images–
This thesis uses a whole heart model, including LV RV and up to the basal area, for the functional analysis of heart motion, based on a blended parameterized deformable model, which is generic enough to deal with different hearts.
- Modeling and analysis of the RV and LV from MRI-SPAMM–
Methods for modeling the RV geometry, extracting RV motion from MRI images using a computer vision/image analysis technique called deformable modeling, and analyzing these results using techniques from mechanics are described.
- –
- Coupled models for visualizing respiratory mechanics–
3D anatomical modeling, involving kinematics and deformable body dynamics, gives a qualitative perspective on the anatomical changes, and a 3D simulation environment for modeling respiratory mechanics and certain cardiopulmonary interactions is created.
- –
- Recursive Dynamics and Adaptive Control for Animating Articulated Figures–
The method presented provides the direct control of kinematic animation to dynamically generated motion, hence taking advantage of the strengths of both techniques.
- –
- Volumetric Analysis of the Left Ventricular Wall Motion from MRI-SPAMM–
Estimating the volumetric shape, motion and deformations of the left ventricle (LV) accurately, and in a clinically useful way, is a very important yet open research problem.
- Integration of Qualitative Segmentation and PBM Methods–
This work uses the qualitative correspondence of edge segments in the images and related contours on the models to provide strong fitting constraints to the physics-based estimation technique presented in the previous chapters.
- –
- –
- –
- Reverse Engineering using Blending and Adaptive Shape Evolution–
A method for reverse engineering that is based on the use of topologically adaptive deformable models based on a shape representation scheme which allows two shapes to be combined into a single model, making the recovered shape a natural symbolic description.
- Multi-Level Shape Representation–
The surface shape is estimated based on the parameters of a superquadric that is subjected to global deformations (tapering and bending) and a variable number of levels of local deformations.
- pe and Motion Esti–
A detailed face model allows for the handling of self-occlusion and the correction of error accumulation in tracking by aligning the expected location of face features with those found in the input images.
- –
- Fast Dynamic Point-to-Point Constraint Algorithm for Deformable Bodies–
An algorithm for the efficient computation of constraint forces necessary for the modeling of hard point-to-point constraints in deformable multibody objects is devised based on a stabilized Lagrange multiplier technique.
- Physics-Based Modeling of Nonrigid Objects for Vision and Graphics (Dissertation)–
This thesis develops a physics-ba,sed framework for 3D slia,pe and nonrigid motion modeling for computer vision and computer gra,phics, based on the use of a new class of dynamically deforniable primitives which allow the combination of globa,l and 1oca.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.