Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Suggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works266 from public data
- Aggregated Residual Transformations for Deep Neural Networks11,971
On the ImageNet-1K dataset, it is empirically show that even under the restricted condition of maintaining complexity, increasing cardinality is able to improve classification accuracy and is more effective than going deeper or wider when the authors increase the capacity.
- Holistically-Nested Edge Detection3,991
HED performs image-to-image prediction by means of a deep learning model that leverages fully convolutional neural networks and deeply-supervised nets, and automatically learns rich hierarchical representations that are important in order to resolve the challenging ambiguity in edge and object boundary detection.
- Deeplysupervised nets2,565
The proposed deeply-supervised nets (DSN) method simultaneously minimizes classification error while making the learning process of hidden layers direct and transparent, and extends techniques from stochastic gradient methods to analyze the algorithm.
- Similarity network fusion for aggregating data types on a genomic scale2,022
Similarity network fusion substantially outperforms single data type analysis and established integrative approaches when identifying cancer subtypes and is effective for predicting survival.
- Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification1,527
It is shown that it is possible to replace many of the 3D convolutions by low-cost 2D convolution, suggesting that temporal representation learning on high-level “semantic” features is more useful.
- Deeply Supervised Salient Object Detection with Short Connections1,458
This paper proposes a new salient object detection method by introducing short connections to the skip-layer structures within the HED architecture, which takes full advantage of multi-level and multi-scale features extracted from FCNs, providing more advanced representations at each layer, a property that is critically needed to perform segment detection.
- Integral Channel Features1,350
It is demonstrated that when designed properly, integral channel features not only outperform other features including histogram of oriented gradient (HOG), they also result in fast detectors when coupled with cascade classifiers.
- 879
- 647
- 625
- 600
- Robust Point Matching via Vector Field Consensus566
This paper proposes an efficient algorithm, called vector field consensus, for establishing robust point correspondences between two sets of points, and suggests a two-stage strategy, where the nonparametric model is used to reduce the size of the putative set and a parametric variant of the approach to estimate the geometric parameters.
- Image segmentation by data-driven markov chain monte carlo565
The DDMCMC paradigm provides a unifying framework in which the role of many existing segmentation algorithms are revealed as either realizing Markov chain dynamics or computing importance proposal probabilities and generalizes these segmentation methods in a principled way.
- 507
- 500
- Co-Scale Conv-Attentional Image Transformers481
Co-scale conv-attentional image Transformers (CoaT), a Transformer-based image classifier equipped with co-scale and convolution-like mechanisms, empowers image Transformers with enriched multi- scale and contextual modeling capabilities.
- Cluster-Based Co-Saliency Detection427
This paper introduces a new cluster-based algorithm for co-saliency detection that is mostly bottom-up without heavy learning, and outperforms most the state-of-the-art saliency detection methods.
- 345
- Weakly supervised histopathology cancer image segmentation and classification313
This paper embeds the clustering concept into the multiple instance learning (MIL) setting and derive a principled solution to performing the above three tasks in an integrated framework and introduces contextual constraints as a prior for MCIL, which further reduces the ambiguity in MIL.
- 307
- Pose Recognition with Cascade Transformers292
This paper utilizes the encoder-decoder structure in Transformers to perform regression-based person and keypoint detection that is general-purpose and requires less heuristic design compared with the existing approaches.
- 292
- 280
- MeMOT: Multi-Object Tracking with Memory275
An online tracking algorithm that performs the object detection and data association under a common framework, capable of linking objects after a long time span is proposed, by preserving a large spatio-temporal memory to store the identity embeddings of the tracked objects.
- Rethinking Spatiotemporal Feature Learning For Video Understanding.272
Interestingly, it was found that 3D convolutions at the top layers of the network contribute more than 3D Convolutional networks at the bottom layers, while also being computationally more efficient, indicating that I3D is better at capturing high-level temporal patterns than low-level motion signals.
- Automated mapping of hippocampal atrophy in 1-year repeat MRI data from 490 subjects with Alzheimer's disease, mild cognitive impairment, and elderly controls257
3D hippocampal volume loss rates increased with worsening diagnosis, and correlated with both baseline and interval changes in Mini-Mental State Examination (MMSE) scores and global and sum-of-boxes Clinical Dementia Rating scale (CDR) scores.
- Comparison of AdaBoost and Support Vector Machines for Detecting Alzheimer's Disease Through Automated Hippocampal Segmentation255
The authors' AdaBoost and Ada-SVM segmentations compared favorably with the manual segmentations and detected disease effects as well as FreeSurfer on the data tested, demonstrating how segmentation accuracy depended on the size of the training set, providing practical information for future users of this technique.
- Automated hippocampal shape analysis predicts the onset of dementia in mild cognitive impairment239
It is concluded that automated shape analysis generates sensitive measurements of early neurodegeneration which predates the onset of dementia and thus provides a prognostic biomarker for conversion of MCI to AD.
- Single-Stage Diffusion NeRF: A Unified Approach to 3D Generation and Reconstruction234
SSDNeRF is presented, a unified approach that employs an expressive diffusion model to learn a generalizable prior of neural radiance fields (NeRF) from multi-view images of diverse objects to enable simultaneous 3D reconstruction and prior learning, even from sparsely available views.
- 228
- Automated 3D mapping of hippocampal atrophy and its clinical correlates in 400 subjects with Alzheimer's disease, mild cognitive impairment, and elderly controls226
This work used a new method for automated hippocampal segmentation, called the auto context model, to analyze brain MRI scans of 400 subjects from the Alzheimer's disease neuroimaging initiative, and gradually reduced sample sizes and used false discovery rate curves to examine the method's power to detect associations with diagnosis and cognition in smaller samples.
- Shape Matching and Recognition – Using Generative Models and Informative Features219
This work presents an algorithm for shape matching and recognition based on a generative model for how one shape can be generated by the other and illustrates how this approach can be generalized to a wider range of problems including object detection.
- 213
- 211
- Regularized vector field learning with sparse approximation for mismatch removal193
This paper proposes a sparse approximation to a robust vector field learning method, sparse vector field consensus (SparseVFC), and derives a statistical learning bound on the speed of the convergence, and applies SparseVFC to the mismatch removal problem.
- 182
- Exponential Moving Average Normalization for Self-supervised and Semi-supervised Learning177
A plug-in replacement for batch normalization (BN) called exponential moving averagenormalization (EMAN) is presented, which improves the performance of existing student-teacher based self- and semi-supervised learning techniques.
- 168
- 161
- 160
- 156
- Improving Shape Retrieval by Learning Graph Transduction155
This paper considers the existing shapes as a group, and study their similarity measures to the query shape in a graph structure, and learns a better metric through graph transduction by propagating the model through existing shapes, in a way similar to com- puting geodesics in shape manifold.
- Line Segment Detection Using Transformers without Edges153
A joint end-to-end line segment detection algorithm using Transformers that is post-processing and heuristics-guided intermediate processing (edge/junction/region detection) free and takes advantages of having integrated tokenized queries, a self-attention mechanism, and encoding-decoding strategy within Transformers.
- 153
- 152
- Multiple Component Learning for Object Detection152
The method, Multiple Component Learning (mcl), automatically learns individual component classifiers and combines these into an overall classifier, and unlike methods that are not part-based, mcl is quite robust to occlusions.
- 150
- Text Spotting Transformers144
This paper presents TExt Spotting TRansformers (TESTR), a generic end-to-end text spotting framework using Transformers for text detection and recognition in the wild, and designs a bounding-box guided polygon detection process.
- 135
- 133
- Guided Variational Autoencoder for Disentanglement Learning130
An algorithm, guided variational autoencoder (Guided-VAE), that is able to learn a controllable generative model by performing latent representation disentanglement learning by providing signal to the latent encoding/embedding in VAE without changing its main backbone architecture.
- 124
- 123
- 123
- Max-Margin Multiple-Instance Dictionary Learning117
This paper designs a formulation to simultaneously learn mixtures of codes by maximizing classification margins in MIL, and results are observed in image classification benchmarks based on the learned codebooks, which observe both compactness and effectiveness.
- DiffusionRig: Learning Personalized Priors for Facial Appearance Editing113
Qualitative and quantitative experiments show that DiffusionRig outperforms existing approaches in both identity preservation and photorealism.
- 107
- Simultaneous Learning and Alignment: Multi-Instance and Multi-Pose Learning ?107
A unified boosting framework for simultaneous learn- ing and alignment is presented and a novel boosting algorithm for Multiple Pose Learning (mpl) is presented, where the goal is to simultaneously split data into groups and train classifiers for each.
- 105
- 102
- Dual Contradistinctive Generative Autoencoder97
The two contradistinctive losses in VAE work harmoniously in DC-VAE leading to a significant qualitative and quantitative performance enhancement over the baseline VAEs without architectural changes.
- Open-Vocabulary Universal Image Segmentation with MaskCLIP96
A Relative Mask Attention (RMA) module is designed to account for segmentations as additional tokens to the ViT CLIP model to perform semantic segmentation and object instance segmentation in open-vocabulary panoptic segmentation.
- 95
- 94
- Dense Volume-to-Volume Vascular Boundary Detection90
This work introduces HED-3D, a 3D extension of the state-of-the-art 2D edge detector (HED), and develops a novel 3D-Convolutional Neural Network (CNN) architecture, I2I- 3D, that predicts boundary location in volumetric data.
- 89
- 89
- 80
- 80
- Learning Instance Occlusion for Panoptic Segmentation79
This work proposes a branch that is tasked with modeling how two instance masks should overlap one another as a binary relation and obtains state-of-the-art results on COCO and show competitive results on the Cityscapes panoptic segmentation benchmark.
- Combining Generative and Discriminative Models for Semantic Segmentation of CT Scans via Active Learning79
This paper presents a new supervised learning framework for the efficient recognition and segmentation of anatomical structures in 3D computed tomography (CT), with as little training data as possible, and a combined generative-discriminative model which increases segmentation accuracy.
- 78
- 77
- 77
- AffordanceLLM: Grounding Affordance from Vision Language Models75
This paper makes an attempt to improve the generalization capability of the current affordance grounding by taking the advantage of the rich world, abstract, and human-object-interaction knowledge from pre-trained large-scale vision language models.
- One-Pixel Signature: Characterizing CNN Models for Backdoor Detection73
The proposed one-pixel signature is a general representation that can be used to characterize CNN models beyond backdoor detection and demonstrates a substantial improvement over the existing competing methods for backdoored CNN detection/classification.
- 72
- Deep Convolutional Neural Networks with Merge-and-Run Mappings70
This work shows that the merge-and-run mapping is a linear idempotent function in which the transformation matrix is idem Potent, and thus improves information flow, making training easy.
- HFS: Hierarchical Feature Selection for Efficient Image Segmentation69
This paper proposes a real-time system, Hierarchical Feature Selection (HFS), that performs image segmentation at a speed of 50 frames-per-second, and can be used in a variety computer vision tasks that are built on top ofimage segmentation and superpixel extraction.
- 68
- Direct mapping of hippocampal surfaces with intrinsic shape context68
A new method for the mapping of hippocampal (HC) surfaces to establish correspondences between points on HC surfaces and enable localized HC shape analysis and shows that the AD group has a different trend in temporal changes of HC asymmetry than the group of normal controls.
- 66
- Identifying Heritable Brain Phenotypes in an Extended Pedigree of Vervet Monkeys65
MRI scanning of monkeys from the extended multigenerational pedigree of the Vervet Research Colony indicates that additive genetic factors account for almost all of the observed variance in brain structure, and suggests the potential of the VRC for genetic mapping of quantitative trait loci underlying such variance.
- 64
- Patched Denoising Diffusion Models For High-Resolution Image Synthesis62
The algorithm Patch-DM is named, in which a new feature collage strategy is designed to avoid the boundary artifact when synthesizing large-size images, and reduces memory complexity compared to the classic diffusion models.
- Wasserstein Introspective Neural Networks62
Wasserstein introspective neural networks (WINN) is presented that are both a generator and a discriminator within a single model that enhances INN's generative modeling capability and gives rise to improved robustness against adversarial examples in terms of the error reduction.
- Learning to predict from crowdsourced data61
A novel mixture model is employed for worker annotations, which learns a prediction model directly from samples to labels for efficient out-of-sample testing.
- 55
- Automatic Subcortical Segmentation Using a Contextual Model53
This work proposes an automatic subcortical segmentation algorithm using the auto context model and develops a framework based on machine learning to learn a unified appearance and context model.
- Parsing Images into Regions, Curves, and Curve Groups53
An algorithm for parsing natural images into middle level vision representations—regions, curves, and curve groups (parallel curves and trees) is presented for an integrated solution to image segmentation and curve grouping through Bayesian inference.
- TokenCompose: Text-to-Image Diffusion with Token-Level Supervision52
TokenCompose aims to improve multi-category instance composition by introducing the token-wise consistency terms between the image content and object segmentation maps in the finetuning stage of the Latent Diffusion Model.
- 52
- Distilling Large Vision-Language Model with Out-of-Distribution Generalizability51
This paper investigates the distillation of visual representations in large teacher vision-language models into lightweight student models using a small- or mid-scale dataset and proposes two principles from vision and language modality perspectives to enhance student’s OOD generalization.
- 51
- 49
- HOIDiffusion: Generating Realistic 3D Hand-Object Interaction Data46
This model is a conditional diffusion model that takes both the 3D hand-object geometric structure and text description as inputs for image synthesis, which offers a more control-lable and realistic synthesis as it can specify the structure and style inputs in a disentangled manner.
- On the Scalability of Diffusion-based Text-to-Image Generation45
This work empirically study the scaling properties of diffusion based T2I models by performing extensive and rigours ablations on scaling both denoising backbones and training set, including training scaled UNet and Transformer variants.
- Visual Relationship Detection Using Part-and-Sum Transformers with Composite Queries45
This paper presents a new approach, denoted Part-and-Sum detection Transformer (PST), to perform end-to-end visual composite set detection and reports experiments on two vision tasks, visual relationship detection and human object interaction and demonstrates that PST achieves state of the art results among single- stage models, while nearly matching the results of custom designed two-stage models.
- 45
- 42
- 41
- Context-Constrained Multiple Instance Learning for Histopathology Image Segmentation41
A new algorithm along the line of weakly supervised learning is proposed; context constraints as a prior for multiple instance learning (ccMIL) is introduced, which significantly reduces the ambiguity in weak supervision (a 20% gain).
- 40
- 39
- Top-Down Learning for Structured Labeling with Convolutional Pseudoprior39
A new method for structured labeling by developing convolutional pseudo-prior (ConvPP) on the ground-truth labels using pseudo-likelihood approximation to the prior under a novel fixed-point network structure that facilitates an end-to-end learning process.
- 38
- Shape matching and registration by data-driven EM38
An efficient and robust algorithm for shape matching, registration, and detection which combines the generative and discriminative models and is very fast, due to the effectiveness of shape-features for solving correspondence requiring typically only four iterations.
- DocTr: Document Transformer for Structured Information Extraction in Documents37
A new formulation for structured information extraction from visually rich documents is presented, motivated by anchor-based object detectors in computer vision, that represents an entity as an anchor word and a bounding box, and represents entity linking as the association between anchor words.
- 37
- Fixed-Point Model For Structured Labeling37
An algorithm with a new perspective on layered models; it aims to find a fixed-point function with the structured labels being both the output and the input, which alleviates the burden in learning multiple/different classifiers in different layers.
- On the Connection of Deep Fusion to Ensembling.36
A new deeply-fused network is developed that combines two networks in a merge-and-run fusion manner that is less deep than a ResNet but yields an ensemble of the same number of more-capable component networks, thus improving the classification accuracy.
- Robust Skull Stripping of Clinical Glioblastoma Multiforme Data34
The ROBEX skull stripping method is extended, using a shape model trained on healthy brains to be relatively insensitive to lesions inside the brain and showing significant improvement over three popular skull stripping algorithms in a dataset of 48 GBM cases.
- Characterizing Learning Curves During Language Model Pre-Training: Learning, Forgetting, and Stability33
It is argued for the existence of sequential learning dependencies between different model capabilities, and language model learning is characterized as early n-gram learning before gradual refinement of tail n-gram predictions.
- Joint Sulcal Detection on Cortical Surfaces With Graphical Models and Boosted Priors32
An automated approach for the joint detection of major sulci on cortical surfaces by representing sulci as nodes in a graphical model that incorporates Markovian relations between sulci and forms their detection as a maximum a posteriori (MAP) estimation problem over the joint space ofmajor sulci.
- Image parsing: segmentation, detection, and recognition32
This method makes use of bottom-up proposals combined with top-down generative models using the Data Driven Markov Chain Monte Carlo (DDMCMC) algorithm which is guaranteed to converge to the optimal estimate asymptotically.
- Reinforcement-aware Knowledge Distillation for LLM Reasoning31
RL-aware distillation (RLAD), which performs selective imitation during RL -- guiding the student toward the teacher only when it improves the current policy update, consistently outperforms offline distillation, standard GRPO, and KL-based on-policy teacher-student knowledge distillation.
- Compatibility-aware Heterogeneous Visual Search31
Compared to ordinary visual search using the largest embedding model (paragon), Cmp-NAS achieves 80-fold and 23-fold cost reduction while maintaining accuracy within 0.3% and 1.6% of the paragon on DeepFashion2 and IJB-C respectively.
- Rotation-Invariant Features for Multi-Oriented Text Detection in Natural Images31
A new algorithm to detect texts of varying orientations is proposed, based on a two-level classification scheme and two sets of features specially designed for capturing the intrinsic characteristics of texts, which achieves significantly enhanced performance on variant texts in complex natural scenes.
- A Learning Based Approach for 3D Segmentation and Colon Detagging31
This paper proposes a new learning based approach for 3D segmentation, and shows its application on colon detagging, and devise a 3D topology representation, grid-line, to perform fast boundary evolution.
- OmniControlNet: Dual-stage Integration for Conditional Image Generation30
This work provides a two-way integration for the widely adopted ControlNet by integrating external condition generation algorithms into a single dense prediction method and incorporating its individually trained image generation processes into a single model.
- Bayesian Diffusion Models for 3D Shape Reconstruction30
The specialty of the BDM lies in its capability to engage the active and effective information exchange and fusion of the top-down and bottom-up processes where each itself is a diffusion process.
- 29
- Automatic Segmentation of MS Lesions Using a Contextual Model for the MICCAI Grand Challenge29
A framework based on machine learning is developed to learn a unified appearance and context model for automatic subcortical segmentation in brain images using the auto context model.
- 27
- 3D Volumetric Modeling with Introspective Neural Networks27
3DWINN is a general framework which enjoys the same properties as WINN in the 2D case: being simultaneously generative and discriminative and it can be applied to the emerging tasks for 3D object and scene modeling.
- Segmentation of Sub-cortical Structures by the Graph-Shifts Algorithm27
The graphshifts algorithm makes use of a dynamic hierarchical representation of the image that allows each iteration of the algorithm to make both small and large changes in the segmentation, similar to PDE and split-and-merge methods, respectively.
- 26
- 26
- 26
- DepR: Depth Guided Single-View Scene Reconstruction with Instance-Level Diffusion25
DepR is proposed, a depth-guided single-view scene reconstruction framework that integrates instance-level diffusion within a compositional paradigm and introduces depth-guided conditioning to effectively encode shape priors into diffusion models.
- Binarized Convolutional Neural Networks with Separable Filters for Efficient Hardware Acceleration25
This paper proposes BCNN with Separable Filters (BCNNw/SF), which applies Singular Value Decomposition (SVD) on BCNN kernels to further reduce computational and storage complexity.
- A classification approach to coreference in discharge summaries: 2011 i2b2 challenge25
This study demonstrates that it is feasible to accomplish the coreference task in discharge summaries for the 2011 i2b2 challenge and achieved encouraging results.
- 25
- 24
- Parsing Images into Region and Curve Processes24
A stochastic algorithm for parsing an image into a number of region and curve processes that achieves global optimization by effective Markov chain Monte Carlo methods in the sense of maximizing a posterior probability.
- 23
- Classification of Alzheimer’s Disease Using a Self-Smoothing Operator21
This system is able to learn/fuse registration-based and overlap-based similarity measures, which are enhanced using a self-smoothing operator (SSO), and shows significantly improved accuracy for Alzheimer's Disease over Diffusion Maps and a popular metric learning approach.
- 20
- 20
- 20
- On the Feasibility of Cross-Task Transfer with Model-Based Reinforcement Learning19
This work proposes Model-Based Cross-Task Transfer (XTRA), a framework for sample-efficient online RL with scalable pretraining and finetuning of learned world models, and achieves substantial improvements over a baseline trained from scratch.
- Convolutions and Self-Attention: Re-interpreting Relative Positions in Pre-trained Language Models19
Composite attention is proposed, which unites previous relative position encoding methods under a convolutional framework, and finds that convolutions consistently improve performance on multiple downstream tasks, replacing absolute position embeddings.
- Robust Subspace Discovery via Relaxed Rank Minimization19
An efficient and effective algorithm based on the alternating direction method of multipliers is given and a new formulation to simultaneously infer the label information and learn the model using low-rank optimization is proposed.
- AUTOMATIC RECOGNITION OF CIVIL INFRASTRUCTURE OBJECTS IN MOBILE MAPPING IMAGERY USING A MARKOV RANDOM FIELD MODEL19
A novel system that generates hot-spot maps using colorindexing and edge gradient indexing and recognizes traffic lights using MCMC (Markov Chain Monte Carlo) method is proposed and is much faster than general color image segmentation and thus is practical to be applied in a recognition system.
- Enhancing Vision-Language Pre-Training with Rich Supervisions18
It is demonstrated that, compared to current screenshot pre-training objectives, this innovative pre-training method significantly enhances performance of image-to-text model in nine varied and popular downstream tasks - up to 76.1% improvements on Table Detection, and at least 1 % on Widget Captioning.
- Local Binary Pattern Networks17
A novel approach to character recognition using local binary pattern networks, or LBPNet, that can learn and perform bit-wise operations in an end-to-end fashion, providing an important means to improve memory efficiency as well as inference speed.
- 16
- 15
- Object-Centric Multiple Object Tracking15
This paper proposes a video object-centric model for MOT that consists of an index-merge module that adapts the object-centric slots into detection outputs and an object memory module that builds complete object prototypes to handle occlusions.
- Geometry-Aware End-to-End Skeleton Detection.15
A new objective function inspired by the Hausdorff distance is introduced that carries both global and local shape information and is made differentiable through an end-to-end neural network framework.
- 15
- Object Recognition Using Junctions15
An object detection/recognition algorithm based on a new set of shape-driven features and morphological operators that is robust to a certain degree of scale change and has advantages of recognizing object parts and dealing with occlusions.
- Structural Edge Detection for Cardiovascular Modeling14
This paper proposes a novel approach to medical image edge detection by adopting the powerful structured forest detector and extending its application to the medical imaging domain and directly incorporates an adaptive prior to create a robust three-dimensional edge classifier.
- Machine Learning for Brain Image Segmentation14
This chapter reviews a variety of algorithms developed by different groups for automatically segmenting structures in medical images, such as brain MRI scans, and presents several different learning techniques increasing in complexity until two algorithms recently proposed by the authors are derived.
- Medical Computer Vision. Recognition Techniques and Applications in Medical Imaging13
Test results showed that REALMS can localize the tumor in 10.89 ms (91.82 fps) with 2.56 ± 1.11 mm errors using a single projection image, showing REALMS’s high potential to support realtime, accurate, and low-dose IGRT.
- Agreement-Based Semi-supervised Learning for Skull Stripping13
A semi-supervised learning algorithm is proposed and its application to skull stripping in brain MRI is shown, taking advantage of existing state-of-the-art systems, such as BET and FreeSurfer, to sample unlabeled data in an agreement-based framework.
- Synthetic MRI Signal Standardization: Application to Multi-atlas Analysis13
This work proposes using multi-spectral data to create synthetic MRI scans matched to the intensity distribution of a given dataset using a physical model of acquisition, significantly improving the results obtained with other intensity correction methods.
- Joint Sulci Detection Using Graphical Models and Boosted Priors13
An automated approach for joint sulci detection on cortical surfaces by using graphical models and boosting techniques to incorporate shape priors of major sulci and their Markovian relations, which is especially powerful in capturing the neighboring relations between sulcal lines, which are otherwise hard to be captured by generative models.
- VideoNSA: Native Sparse Attention Scales Video Understanding12
This work adapts Native Sparse Attention to video-language models, employing a hardware-aware hybrid approach to attention, preserving dense attention for text, while employing NSA for video, and achieves improved performance on long-video understanding, temporal reasoning, and spatial benchmarks.
- Restoration by Generation with Constrained Priors12
This work proposes a method to adapt a pretrained diffusion model for image restoration by simply adding noise to the input image to be restored and then denoise, based on the observation that the space of a generative model needs to be constrained.
- 12
- Deformable Templates Guided Discriminative Models for Robust 3D Brain MRI Segmentation12
This paper proposes a robust brain image segmentation algorithm that fuses together deformable templates and informative features and achieves both robustness and efficiency, and can be used to segment brain MRI images with large anatomical variations.
- Suicide Note Sentiment Classification: A Supervised Approach Augmented by Web Data12
The results indicated that classifying fine-grained sentiments at sentence level is a non-trivial task and it is effective to divide categories into different groups according to their semantic properties.
- 11
- Towards Whole Brain Segmentation by a Hybrid Model11
This paper presents a new brain segmentation algorithm using a hybrid model that handles multi-class patterns more easily than the original probabilistic boosting tree (PBT), and it facilitates the process, eventually, toward wholebrain segmentation.
- SG-VLA: Learning Spatially-Grounded Vision-Language-Action Models for Mobile Manipulation10
This work introduces a framework for learning spatially-grounded VLA models that strengthens perception and representation through auxiliary task co-training and multi-modal input enhancement, and suggests that spatial grounding through auxiliary and multi-modal learning provides a strong direction for scaling VLA models toward general-purpose domestic robots.
- 10
- Reverse Image Segmentation: A High-Level Solution to a Low-Level Task10
An interesting observation is presented that performs image segmentation in a reverse way, i.e., using a high-level semantic labeling approach to address a low-level segmentation problem, could be a proper solution.
- Disagreement-Based Multi-system Tracking10
This paper tackles the tracking problem from a fusion angle and proposes a disagreement-based approach, which can be built on top of nearly any existing tracking systems by exploiting their disagreements.
- Modeling diffusion‐weighted MRI as a spatially variant Gaussian mixture: Application to image denoising10
The presented spatially variant mixture model for diffusion MRI provides excellent denoising results at low signal-to-noise ratios and can be fitted to a noisy diffusion MRI volume by maximizing the posterior probability of the underlying noiseless data in a Bayesian framework.
- Convolutional Pseudo-Prior for Structured Labeling.9
A new method for structured labeling by developing convolutional pseudo-prior (ConvPP) on the ground-truth labels using pseudo-likelihood approximation to the prior under a novel fixed-point network structure that facilitates an end-to-end learning process.
- A Learning Based Algorithm for Automatic Extraction of the Cortical Sulci9
A learning based method for automatic extraction of the major cortical sulci from MRI volumes or extracted surfaces using the Probabilistic Boosting Tree algorithm, which implicitly discovers and combines rules based on manually annotated sulci traced by neuroanatomists.
- CVP: Central-Peripheral Vision-Inspired Multimodal Model for Spatial Reasoning8
Two complementary components are introduced into a Large Multimodal Model-based architecture: target-affinity token, analogous to central vision, that guides the model’s attention toward query-relevant objects; and allocentric grid, akin to peripheral vision, that captures global scene context and spatial arrangements to enable structured, context-aware understanding of complex 3D environments.
- Elodi: Ensemble Logit Difference Inhibition for Positive-Congruent Training8
The role of ensembles in reducing NFR is analyzed and it is observed that they remove negative flips that are typically not close to the decision boundary, but often exhibit large deviations in the distance among their logits.
- Lay-Your-Scene: Natural Scene Layout Generation with Diffusion Transformers7
This work proposes to use lightweight open-source language models to obtain scene elements from text prompts and a novel aspect-aware diffusion Transformer architecture trained in an open-vocabulary manner for conditional layout generation.
- Grounded Compositional and Diverse Text-to-3D with Pretrained Multi-View Diffusion Model7
The Grounded-Dreamer method consistently outperforms previous state-of-the-art (SOTA) methods in generating compositional 3D assets, excelling in both quality and accuracy, and enabling diverse 3D from the same text prompt.
- Action-Gons: Action Recognition with a Discriminative Dictionary of Structured Elements with Varying Granularity7
This paper proposes a systematic learning process to build an over-complete dictionary of “Action-Gons”, a middle level representation for action recognition in videos inspired by the Julesz school of n-gon statistics.
- Contexts-Constrained Multiple Instance Learning for Histopathology Image Analysis7
A new algorithm along the line of weakly supervised learning is proposed; context constraints as a prior for multiple instance learning (ccMIL) is introduced, which significantly reduces the ambiguity in weak supervision (a 20% gain).
- Ground-V: Teaching VLMs to Ground Complex Instructions in Pixels6
This work addresses five critical real-world challenges in text-instruction-based grounding: hallucinated references, multi-object scenarios, reasoning, multi-granularity, and part-level references, and generates high-quality instruction-response pairs linked to existing pixel-level annotations, minimizing the need for costly human annotation.
- C3Editor: Achieving Controllable Consistency in 2D Model for 3D Editing6
This work proposes C3Editor, a controllable and consistent 2D-lifting-based 3D editing framework that delivers more consistent and controllable 2D and 3D editing results than existing 2D-lifting-based methods, outperforming them in both qualitative and quantitative evaluations.
- Efficient Scaling of Diffusion Transformers for Text-to-Image Generation6
U-ViT, a pure self-attention based DiT model provides a simpler design and scales more effectively in comparison with cross-attention based DiT variants, which allows straightforward expansion for extra conditions and other modalities.
- 6
- 6
- 6
- Wavelet-Based Representation of Biological Shapes6
The wavelet-space-induced shape representation allows us to study the multi-spectral nature of the shape's geometry, topology and features and is more compact and allows utilization of diverse wavelet bases.
- Non-autoregressive Sequence-to-Sequence Vision-Language Models5
A parallel decoding sequence-to-sequence vision-language model, trained with a Query-CTC loss, that marginalizes over multiple inference paths in the decoder, which achieves performance on-par with its state-of-the-art autoregressive counterpart, but is faster at inference time.
- 5
- 5
- 5
- One-Class Multiple Instance Learning via Robust PCA for Common Object Discovery5
This paper proposes an algorithm using the robust PCA (RPCA) in an iterative way to perform simultaneous common object discovery and model learning under a one-class multiple instance learning setting, which achieves comparable results with standard two-class MIL learning algorithms.
- 5
- A FRAMEWORK FOR AUTOMATIC RECOGNITION OF SPATIAL FEATURES FROM MOBILE MAPPING IMAGERY5
A new framework for object recognition by hypothesis-and-test techniques is proposed and developed and an example of traffic-light recognition from mobile mapping images is given in detail.
- AuthGuard: Generalizable Deepfake Detection via Language Guidance4
The resulting framework, AuthGuard, achieves state-of-the-art deepfake detection accuracy in both in-distribution and out-of-distribution settings, achieving AUC gains of 6.15% on the DFDC dataset and 16.68% on the DF40 dataset.
- 4
- 4
- 4
- Discriminative Clustering via Generative Feature Mapping4
This paper takes the advantages of both models of generative and discriminative clustering by coupling the two paradigms through feature mapping derived from linearizing Bayesian classifiers, and proposes the unified probabilistic framework.
- Fast Approximate Stochastic Tractography4
The proposed FAST algorithm achieves a 60x speed-up over the Monte Carlo sampling scheme, therefore enabling interactive probabilistic tractography: the user can quickly modify the seed region if he is not satisfied with the output without having to wait on average 7 min.
- CyCLeGen: Cycle-Consistent Layout Prediction and Image Generation in Vision Foundation Models3
CyCLeGen is presented, a unified vision-language foundation model capable of both image understanding and image generation within a single autoregressive framework, allowing self-improvement via synthetic supervision under a reinforcement learning objective guided by cycle consistency.
- PixARMesh: Autoregressive Mesh-Native Single-View Scene Reconstruction3
PixARMesh achieves state-of-the-art reconstruction quality while producing lightweight, high-quality meshes ready for downstream applications, and augment a point-cloud encoder with pixel-aligned image features and global scene context via cross-attention, enabling accurate spatial reasoning from a single image.
- Talk2Move: Reinforcement Learning for Text-Instructed Object-Level Geometric Transformation in Scenes3
Talk2Move, a reinforcement learning (RL) based diffusion framework for text-instructed spatial transformation of objects within scenes achieves precise, consistent, and semantically faithful object transformations, outperforming existing text-guided editing approaches in both spatial accuracy and scene coherence.
- 3
- 3
- 3
- 3
- 2
- 2
- Point Cloud Recognition with Position-to-Structure Attention Transformers2
Two new aspects in PS-Former are introduced: a learnable condensation layer that performs point downsampling and feature extraction; and a Position-to-Structure Attention mechanism that recursively enriches the structural information with the position attention branch.
- 2
- 2
- 2
- 2
- Introspective Classifier Learning: Empower Generatively.2
In introspective classifier learning (ICL) that emphasizes the importance of having a discriminative classifier empowered with generative capabilities, a reclassification-by-synthesis algorithm is developed to perform training using a formulation stemmed from the Bayes theory.
- 2
- Overview of the 2013 Workshop on Medical Computer Vision (MCV 2013)2
The workshop aims at exploring the use of modern computer vision technology in tasks such as automatic segmentation and registration, localisation of anatomical features and detection of anomalies, and especially in modern, scalable and efficient algorithms which generalise well to previously unseen images.
- 2
- 2
- MRF Labeling with a Graph-Shifts Algorithm - eScholarship2
An adaptation of the recently proposed graph-shifts algorithm for labeling MRF problems from low-level vision with the potential for such a hierarchical energy minimization algorithm on low- level vision problems with large label sets is presented.
- 2
- Lossless Online Ensemble Learning (LOEL) and Its Application to Subcortical Segmentation2
Among those tested, LOEL largely outperformed the alternative online learning algorithms and gave excellent error metrics that were consistent between the online and offline case; it also accurately distinguished AD subjects from healthy controls based on automated measures of hippocampal volume.
- MRF Labeling with a Graph-Shifts Algorithm2
An adaptation of the recently proposed graph-shifts algorithm for labeling MRF problems from low-level vision with the potential for such a hierarchical energy minimization algorithm on low- level vision problems with large label sets is presented.
- 2
- SemLayer: Semantic-aware Generative Segmentation and Layer Construction for Abstract Icons1
SemLayer is introduced, a visual generation empowered pipeline that restores editable layered structures and enables editing workflows previously inapplicable to flattened vector graphics and establishes semantic layer reconstruction as a practical and valuable task.
- An Integrative Single-Cell Epigenomic Atlas Annotates the Regulatory Genome of the Adult Mouse Brain1
A single-cell epigenome atlas of transcription and four histone modifications and the transcriptome in 2.5 million nuclei from the adult mouse brain is constructed, providing a comprehensive framework for annotating the brain regulatory genome and interpreting non-coding disease risk variants.
- 1
- Real Deep Research for AI, Robotics and Beyond1
This work presents Real Deep Research (RDR) a comprehensive framework applied to the domains of AI and robotics, with a particular focus on foundation models and robotics advancements.
- 1
- SkeleTR: Towrads Skeleton-based Action Recognition in the Wild1
SkeleTR can be directly applied to multiple skeleton-based action tasks, including video-level action classification, instance- level action detection, and group-level activity recognition, and enables transfer learning and joint training across different action tasks and datasets, which result in performance improvement.
- 1
- 1
- Top-down Flow Transformer Networks.1
A comprehensive study on various datasets including MNIST, shapes, and natural images with both inner and inter datasets demonstrates the advantages of the proposed top-down flow transformer framework, which can be adopted in a variety of computer vision applications.
- 1
- A two-layer structure prediction framework for microscopy cell detection1
A new two- layer model for cell centre detection by a two-layer structure prediction framework, which is respectively built on classification for the cell centres implicitly using rich appearances and contextual information and explicit structural information for the cells.
- 1
- 1
- 1
- Learning Context-Sensitive Shape Similarity1
A new perspective to this problem is provided by considering the existing shapes as a group, and study their similarity measures to the query shape in a graph structure, which achieves promising improvements on both shape classification and shape clustering.
- 1
- Automated Hippocampal Segmentation and Mapping Reveals Genetically Accelerated Tissue Loss in 1-year Repeat MRI data from 490 Alzheimer's Disease, MCI, and Control Subjects1
Alzheimer's Disease, MCI, and Control Subjects J H Morra, Z Tu, L G Apostolova, and control subjects.
- Automatic Caudate Segmentation by Hybrid Generative/Discriminative Models1
This paper presents an experimental study for caudate segmentation in MRI images using a fully automatic algorithm based on [9], which is an extension of the hybrid model approach described in [1].
- Effective Bayesian inference by data-driven Markov chain Monte Carlo for object recognition and image segmentation1
The DDMCMC paradigm utilizes data driven techniques, such as Hough transform, edge detection, and color clustering, to design effective transition probabilities for Markov chain dynamics, which drastically improves the effectiveness of traditional MCMC algorithms in terms of two standard metrics: 'burn-in' period and 'mixing' rate.
- –
- –
- –
- Gaussian Swaying ⚑: Surface-Based Framework for Aerodynamic Simulation with 3D Gaussians–
Comprehensive experiments on both synthetic and real-world datasets across multiple metrics demonstrate that Gaussian Swaying achieves state-of-the-art performance and efficiency, offering a scalable approach for realistic aerodynamic scene simulation.
- DLLG: Dynamic Logit-Level Gating of LLM Experts–
DLLG (Dynamic Logit-Level Gating), a dynamic logit-level ensembling framework that learns token-level expert fusion from sparse response-level supervision, highlights learned logit-level fusion as a robust and scalable paradigm for integrating specialized experts.
- –
- –
- LIFT: Layout-In-Future Video Generation under Large Viewpoint Change via On-Policy Self-Distillation–
- –
- –
- –
- –
- –
- Harvesting Motion Patterns in Still Images from the Internet–
It is demonstrated that still images are indeed rich enough for use in motion understanding; in the process, they also characterize those motions that are recognizable even in still images.
- Pseudocontractive Updates: A Common Thread in Convex Optimization Methods–
This manuscript points out and demonstrates the utility of an important but largely unremarked common thread running through many prominent optimization methods, and provides an initial example of the type of unified analysis that becomes possible in the settings where the property of pseudocontractivity is found to hold.
- Medical Computer Vision: Recognition Techniques and Applications in Medical Imaging Second International MICCAI Workshop, MCV 2012, Nice, France, ... Vision, Pattern Recognition, and Graphics)–
The papers explore the use of modern computer vision technology in tasks such as automatic segmentation and registration, localization of anatomical features and detection of anomalies, as well as 3D reconstruction and biophysical model personalization.
- Supplementary Materials for: Unsupervised Object Class Discovery via Saliency-Guided Multiple Class Learning–
This paper does discriminative learning with the presence of hidden variables and estimates the model θ that minimizes the negative log-likelihood function L(θ;Y,X) = − log Pr(Y |X; θ).
- –
- Learning based approaches with applications to vision tasks–
A coarse-to-fine learning-based approach is proposed for image registration which dramatically reduces the search space of possible geometrical transformations and a self-smoothing operator is introduced to improve upon an initial similarity measure which better respects the data manifold structure.
- Fusion deformable template with discriminative models for robust three-dimensional brain image segmentation–
A robust brain image segmentation algorithm is proposed by fusing together deformable templates and informative features by testing the method on a few popular datasets of a large number of images and showing advantages over several existing state-of-the-art systems.
- Shape matching and registration by data-driven EM - eScholarship–
The work in this paper treats shape matching as the best geometrical transformation between two shapes in the spirit of Grenander’s pattern theory and gives a hybrid algorithm which combines the generative and discriminative models.
- –
- Image Parsing by Data-Driven Markov Chain Monte Carlo–
This thesis proposes a stochastic computing paradigm called DDMCMC for visual inference and summarize many traditional vision problems for example, image segmentation, object recognition, perceptual organization etc., into a general problem called image parsing.
- –
- –
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 2026-10-14. 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.