Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks255 from public data
- An overview of multi-modal medical image fusion366
In this review, methods in the field of medical image fusion are characterized by image decomposition and image reconstruction, image fusion rules, image quality assessments, and experiments on the benchmark dataset.
- Shape-Adaptive Selection and Measurement for Oriented Object Detection309
Novel flexible shape-adaptive selection (SA-S) and shape- Adaptive measurement (SA -M) strategies for oriented object detection, which comprise an SA-S strategy for sample selection and SA-M strategy for the quality estimation of positive samples are proposed.
- Representative Discovery of Structure Cues for Weakly-Supervised Image Segmentation240
Experimental results show that the proposed approach outperforms state-of-the-art weakly-supervised image segmentation methods, on five popular segmentation data sets and performs competitively to the fully- supervised segmentation models.
- TransZero: Attribute-Guided Transformer for Zero-Shot Learning206
This paper proposes an attribute-guided Transformer network to learn the attribute localization for discriminative visual-semantic embedding representations in ZSL, termed TransZero, which achieves a new state-of-the-art on three ZSL benchmarks.
- 201
- Single image dehazing with a physical model and dark channel prior171
The proposed single image dehazing method is based on a physical model and the dark channel prior principle and the selection of an atmospheric light value is directly responsible for the color authenticity and contrast of the resulting image.
- 157
- A Real-Time Hand Posture Recognition System Using Deep Neural Networks148
This article proposes a two-stage HPR system for Sign Language Recognition using a Kinect sensor and applies deep neural networks (DNNs) to automatically learn features from hand posture images that are insensitive to movement, scaling, and rotation.
- 141
- A novel approach of lung segmentation on chest CT images using graph cuts114
A new algorithm based on an improved graph cuts algorithm with Gaussian mixture models (GMMs) that can directly provide explicit lung regions without any postprocessing operations even in complex scenarios is proposed.
- 113
- An improved fractional-order differentiation model for image denoising113
An improved model based on the Grunwald-Letnikov (G-L) fractional differential operator based on G-L fractional order differentiation is proposed for image denoising and successfully used to denoise three-dimensional magnetic resonance images.
- Research on manufacturing grid resource service optimal-selection and composition framework104
The case study and comparison of performances of the algorithms demonstrate that the proposed methods are sound on success rate and executing efficiency.
- G-Rep: Gaussian Representation for Arbitrary-Oriented Object Detection99
A unified Gaussian representation called G-Rep is proposed to construct Gaussian distributions for OBB, QBB, and PointSet, which achieves a unified solution to various representations and problems and shows excellent performance on several publicly available datasets.
- 95
- Performance evaluation of deep feature learning for RGB-D image/video classification94
This paper conducts extensive experiments on five popular RGB-D datasets and presents a detailed analysis about the comparison between the learned feature representations from the four deep learning models.
- Particle Swarm Optimization based dictionary learning for remote sensing big data93
This paper proposes a method of improved online dictionary learning based on Particle Swarm Optimization (PSO), which improves the performance of the algorithm for large-scale remote sensing images, and has a better effect on noise suppression.
- Robust object removal with an exemplar-based image inpainting approach91
In the improved model, a regularized factor is introduced to adjust the patch priority function and a modified sum of squared differences (SSD) and normalized cross correlation are combined to search for the best matching patch.
- 88
- IK-SVD: Dictionary Learning for Spatial Big Data via Incremental Atom Update82
The information, from both old and new samples, is explored in the proposed incremental K-SVD (IK-SVD) algorithm, but only the current atoms are adaptively updated, which makes the dictionary better represent all the samples without the influence of redundant information from old samples.
- 81
- Multiview Hessian regularized logistic regression for action recognition78
This paper carefully conduct extensive experiments on the unstructured social activity attribute (USAA) dataset and the experimental results demonstrate the effectiveness of the proposed multiview Hessian regularized logistic regression for human action recognition.
- 71
- 71
- 70
- Large-scale paralleled sparse principal component analysis69
This paper construct parallel implementations of the four optimization formulations of the generalized power method of SPCA (GP-SPCA), one of the most efficient and effective S PCA approaches, on a GPU.
- 67
- A new perspective: Recognizing online handwritten Chinese characters via 1-dimensional CNN67
This work proposes to apply the 1-dimensional CNN (1-D CNN), which, to their knowledge, is novel in the context of OLHCCR, and attains a faster test time speed compared with the RNN, and this is more pronounced in processing long sequences.
- Learning Correspondence Structures for Person Re-Identification62
This paper addresses the problem of handling spatial misalignments due to camera-view changes or human-pose variations in person re-identification by introducing a boosting-based approach to learn a correspondence structure, which indicates the patchwise matching probabilities between images from a target camera pair and introduces a global constraint-based matching process.
- Data augmentation and directional feature maps extraction for in-air handwritten Chinese character recognition based on convolutional neural network59
CNN is introduced into in-air handwritten Chinese character recognition (IAHCCR), new directional feature maps are proposed and a new data augmentation method dedicated to in- air handwritten Chinese characters is proposed for further improving recognition rate.
- 58
- 55
- 3D model retrieval and classification by semi-supervised learning with content-based similarity54
A novel 3D model retrieval and recognition method that employs both a distance histogram and 3D moment invariants as features that are invariant to 3D object scaling, translation, and rotation is proposed.
- 53
- Group Re-Identification53
Multi-grain representations are computed to characterize the appearance and spatial features of multi-grain objects and evaluate the importance weight of each object for group Re-ID, so as to handle the interferences from group dynamics.
- 49
- Multimodal PEAR Chain-of-Thought Reasoning for Multimodal Sentiment Analysis45
Experimental results on two multimodal sentiment analysis benchmark datasets show that high-level reasoning information can help learn discriminative text representation, and cross-modal complementary information can avoid misleading by unreasonable steps in the CoT.
- Exploring Visual Context for Weakly Supervised Person Search45
This paper inventively considers weakly supervised person search with only bounding box annotations, and discovers three levels of context clues in unconstrained natural images that promote local and global discriminative capabilities and enhances clustering accuracy.
- Parallel programing templates for remote sensing image processing on GPU architectures: design and implementation45
This paper attempts to explore a reusable GPU-based remote sensing image parallel processing model and to establish a set of parallel programming templates, which provides programmers with a more simple and effective way for programming parallelRemote sensing image processing algorithms.
- 42
- Compressing the CNN architecture for in-air handwritten Chinese character recognition42
A unified algorithm to effectively compress the CNN for IAHCCR with little accuracy loss is proposed, which first utilizes the channel pruning strategy to simplify the network structure, and then adopts the network quantization technique to represent parameters with lower precision.
- 3D face reconstruction from skull by regression modeling in shape parameter spaces42
A regression modeling based method is proposed, in which a statistical shape model is built for skulls and faces, respectively, and the relationship between them is extracted in the shape parameter spaces through partial least squares regression (PLSR).
- 3D object detection: Learning 3D bounding boxes from scaled down 2D bounding boxes in RGB-D images41
An efficient 3D object detection system that can predict object location, size, and orientation is presented and outperforms the state-of-the-art detection methods by a remarkable margin with faster detection time.
- 41
- Neighborhood Contrastive Transformer for Change Captioning40
A neighborhood contrastive transformer is proposed to improve the model's perceiving ability for various changes under different scenes and cognition ability for complex syntax structure and introduces the explicit dependencies between words to calibrate the transformer decoder, which helps better understand complex syntax structure during training.
- 39
- 39
- 38
- Action Recognition With Coarse-to-Fine Deep Feature Integration and Asynchronous Fusion38
This paper introduces a coarse-to-fine network which extracts shared deep features at different action class granularities and progressively integrates them to obtain a more accurate feature representation for input actions, and introduces an asynchronous fusion network which fuses information from different streams by asynchronously integrating stream-wise features in different time points.
- ExpLLM: Towards Chain of Thought for Facial Expression Recognition37
A novel method called ExpLLM is proposed, which leverages large language models to generate an accurate chain of thought (CoT) for facial expression recognition, and outperforms current state-of-the-art FER methods.
- 36
- 36
- Nonlocal Means-Based Denoising for Medical Images36
The proposed adaptive NL-means denoising scheme has been compared with similar state-of-art methods over synthetic and real clinical medical images showing an improved performance in all cases analyzed.
- A size-invariant convolutional network with dense connectivity applied to retinal vessel segmentation measured by a unique index35
An ingenious index: fusion score (FS), which provides an overall measure for those binary images of RVS tasks, which outperforms other techniques when evaluated in terms of F1-score, Matthews correlation coefficient, G-mean and FS.
- In-air handwritten Chinese character recognition with locality-sensitive sparse representation toward optimized prototype classifier35
A locality-sensitive sparse representation toward optimized prototype classifier (LSROPC) for IAHCCR, where the learned dictionary can not only preserve local data structures, but also require the reconstruction of a pattern to get as close to the prototype optimized by the minimum classification error (MCE) approach.
- 35
- Social event detection with robust high-order co-clustering33
The MediaEval Social Event Detection Dataset and its subset are selected to demonstrate the effectiveness of the proposed approach in handling the datasets with and without non-event samples.
- Manifold regularized kernel logistic regression for web image annotation31
Manifold regularized kernel logistic regression (KLR) is proposed for web image annotation and has the following advantages: (1) the KLR has a smooth loss function; (2) theKLR produces an explicit estimate of the probability instead of class label; and (3)The KLR can naturally be generalized to the multi-class case.
- 30
- In-air handwritten Chinese text recognition with temporal convolutional recurrent network30
A 3D in-air handwritten Chinese text recognition (IAHCTR) system for the first time is presented, and the first public large-scale IAHCT dataset is constructed, and a novel architecture, named the temporal convolutional recurrent network (TCRN), is proposed for online HCTR.
- 30
- FEAFA: A Well-Annotated Dataset for Facial Expression Analysis and 3D Facial Animation29
A novel method that drives virtual characters only based on action unit value regression of the 2D video frames of source actors is proposed and the potential of FEAFA for 3D facial animation is demonstrated.
- Accelerated nonrigid image registration using improved Levenberg–Marquardt method29
A dedicated optimization method is proposed for nonrigid CT image registration based on Levenberg–Marquardt (L-M) optimization and a linear search for the trial step is introduced to enhance performance.
- 29
- OccCasNet: Occlusion-Aware Cascade Cost Volume for Light Field Depth Estimation27
This work proposes an occlusion-aware cascade cost volume for LF depth (disparity) estimation and introduces occlusion maps to enhance accuracy in constructing the occlusion-aware cost volume, allowing the refined disparity estimation network to yield a more precise disparity map.
- 26
- 25
- 25
- 25
- 23
- 21
- Leveraging 3D blendshape for facial expression recognition using CNN20
In order to train the blendshape network, a large blendshape image dataset is constructed, namely BlendshapeExp, which gets 297,490 images with ground truth blendshape coefficients and is trained to regress these coefficients from a single image.
- GPU-based real-time terrain rendering: Design and implementation20
This paper presents a real time rendering algorithm based on GPU (Graphics Processing Unit) and tessellation technology and proves that the method can highly reduce the processing time and get a feasible result.
- 18
- Compressed sensing based remote sensing image reconstruction via employing similarities of reference images18
An algorithm to reconstruct the target image with information from multi-source and multi-temporal reference images to improve the image reconstruction accuracy and the degree of similarity between the reconstructed image and thetarget image is proposed.
- Content-based similarity for 3D model retrieval and classification17
This paper proposes a mutual information distance measurement to perform the similarity comparison of 3D objects and proposes a method combining a distance histogram and moment invariants to improve the retrieval performance.
- 16
- Adaptively weighted nonlocal means and TV minimization for speckle reduction in SAR images16
The proposed model is intended for the logarithmic domain of SAR data, and combines the above two methods by minimizing an adaptive TV using a nonlocal data fidelity term and a fast iterative shrinkage/thresholding algorithm (FISTA) to solve the optimization problem.
- Guided trilateral filter and its application to ultrasound image despeckling16
A novel local spatial filtering framework named as the guided trilateral filter (GTF) to implement restoration of a noisy image with a well-fitting statistical distribution model is proposed and demonstrated that the GTDF not only can achieve superior performance against state of the art methods, but also has fast and robust convergence and parameter setting insensitivity.
- 16
- Non-local sparse regularization model with application to image denoising16
Experimental results from image denoising tasks of synthetic and real noisy images show that the proposed method outperforms the state-of-the-art, making it possible to effectively restore raw images from digital cameras at a reasonable speed and memory cost.
- A graph distance based metric for data oriented workflow retrieval with variable time constraints16
A graph distance based approach for measuring the similarity between data oriented workflows with variable time constraints is proposed and a formal structure called Time Dependency Graph (TDG) is proposedand further used as representation model of workflows.
- Single image motion deblurring with reduced ringing effects using variational Bayesian estimation15
Experimental results demonstrate that the blind restoration algorithm that is proposed in this paper can effectively remove image blurring caused by camera shake and obtain restored images that were clearer and more reliable.
- Realistic Full-Body Motion Generation from Sparse Tracking with State Space Model14
Motion Mamba Diffusion (MMD) is presented, a novel conditional diffusion model, which effectively utilizes the sequence modeling capability of SSMs and the robust generation ability of diffusion models to track full-body poses accurately, and design a bidirectional Temporal Mamba Module (TMM) to model motion sequence.
- Deep Hashing Based on VAE‐GAN for Efficient Similarity Retrieval14
A VAE-GAN based hashing framework for fast image retrieval that combines a Variational autoencoder (VAE) with a Generative adversarial network (GAN) to generate content preserving images for pairwise hashing learning.
- A flexible 3D cerebrovascular extraction from TOF-MRA images14
A novel active contour model is presented by combining the statistical information and the vessel shape information in a variational level set framework, and the results show that the proposed model outperforms the classical models.
- 13
- 12
- Transmission of multimedia contents in opportunistic networks with social selfish nodes12
In the proposed scheme, priority is given to a message according to the cooperation degree, a cache management strategy is applied, and a delivery probability is calculated to guide nodes to select the better nodes as their relays.
- Kernel Optimization for Blind Motion Deblurring with Image Edge Prior12
A blind motiondeblurring approach that solves blur kernel and the latent image robustly and achieves better deblurring results than previous works is proposed.
- Group-Wise Learning for Aurora Image Classification With Multiple Representations11
This work proposes a group-wise learning (GWL) method for the automatic aurora image classification using multiple representations, which focuses on the diverse properties of multiple feature representations, by clustering the correlated representations into the same group.
- Energy-based automatic recognition of multiple spheres in three-dimensional point cloud11
A novel method for automatic recognition of multiple spheres in 3D point cloud that outperforms the Hough-based and the RANSAC-based method in accuracy and robustness and alleviates the dependence of existing algorithms on distance thresholds, the requirement of the unknown number and parameters of spheres, and the requirements of a huge sampling number to generate initial models.
- 11
- An adaptively weighted algorithm for camera calibration with 1D objects11
Experiments with synthetic and real image data show that the accuracy of the proposed algorithm is much higher than that of classical 1D calibration and in some cases, it outperforms the nonlinear optimal 1D algorithm and is even comparable with the normalized 2D and 3D calibration.
- 10
- 10
- SPN: short path network for scene text detection10
A novel text detector, named short path network (SPN), with superior performance in both accuracy and efficiency, which achieves the state-of-the-art performance on three scene text benchmarks, in terms of F-score and frame per second (FPS).
- Color Image Denoising Based on Guided Filter and Adaptive Wavelet Threshold10
A method that further enhances the edge and texture details of the image using guided filtering and achieves superior color image denoising compared to these conventional techniques is described.
- Convex optimization based low-rank matrix decomposition for image restoration10
The method seeks an optimal set of image domain transformations such that the matrix of transformed images can be decomposed as the sum of a sparse matrix of errors and a low-rank matrix of recovered denoised images, which can be solved efficiently using scalable convex optimization techniques.
- Fields of Experts Based Multichannel Compressed Sensing10
The experimental results show that the new compressed sensing-based model and an efficient algorithm for image reconstruction from sparse observations has competitive results for single-band images and outweighs all the other methods for color images, and so it is an appropriate model for reconstructing multichannel images.
- YOLO-mini-tiger: Amur Tiger Detection9
An efficient deep tiger detector, which consists of the convnet channel adaptation method and an improved tiger detection method based on You Only Look Once version 3, which outperforms previous Amur tiger detection methods presented at CVWC2019.
- G-IK-SVD: parallel IK-SVD on GPUs for sparse representation of spatial big data9
The contemporary general-purpose computing on the graphics processing unit (GPGPU) is employed to extend a sparse representation method for big image datasets, IK-SVD, namely G-IK-SVA, to show results that indicate that the GPU-aided batch-OMP algorithm shows speedups of up to 30 times than the sparse coding part of IK -SVD.
- 9
- 9
- 8
- 8
- Defect Detection Based on Improved YOLOx for Ultrasonic Images8
An improved object detection method for defect detection that adopts the four-scale Swin Transformer as the multi-scale feature extraction network and uses the decoupled head from YOLOx to output defect categories and locations is proposed.
- 8
- 8
- A 3D polar-radius-moment invariant as a shape circularity measure8
The experimental results show that the proposed circularity measure can accord with human visual perception and consistently performs best in terms of retrieval efficiency in test datasets than other compared methods.
- 8
- 8
- 7
- 7
- 7
- Auto-Segmentation of Pathological Lung Parenchyma Based on Region Growing Method7
The results show that the improved convex hull algorithm can repair the concavities of lung contour effectively and the proposed segmentation method can extract the lung parenchyma precisely.
- Efficient volume rendering methods for out-of-Core datasets by semi-adaptive partitioning7
A semi-adaptive partitioning strategy and an efficient out-of-core visualization framework with improved volume rendering algorithms are described and Experimental results indicate that the new methods are effective and efficient for visualizing out- of-core 3D scalar field datasets.
- A Method of Protein Model Classification and Retrieval Using Bag-of-Visual-Features7
Experimental results show that the proposed method has encouraging performances for protein retrieval and categorization as shown in the comparison with other methods.
- Single-image motion deblurring using an adaptive image prior7
By identifying the correct image prior for each pixel in the image, this approach models the spatially varying motion blur exhibited by camera motion more effectively than conventional methods based on space-invariant blur kernels.
- An efficient multi-threshold AdaBoost approach to detecting faces in images7
The experimental results show that the usage of the proposed multi-threshold weak classifier leads to an effective face detector with much less features and at the same time maintaining the same or even better performance, when comparing to the conventional single-threshold weak classifier.
- 6
- R-FENet: A Region-based Facial Expression Recognition Method Inspired by Semantic Information of Action Units6
A region-based convolutional neural network for the recognition of facial expression named R-FENet that achieves state-of-the-art results with accuracy of 60.95% on AffectNet and 55.97% on SFEW, relative to single-model methods.
- 6
- Gait, Stability and Movement of Snake-Like Robots6
Stability, crawling gait, moving velocity, climbing capability, the capability to cross ditches and avoid obstacles, etc, are analyzed.
- Images Denoising by Improved Non-Local Means Algorithm6
This work will implement the non-local means algorithm addressing the preservation of structure in a digital image and compare it to other denoising methods in experimental results.
- 5
- 5
- Human-object interaction detection via interactive visual-semantic graph learning5
This work fuse the contextual features from the visual graph and semantic graph with the visual characteristics of the human-object pairs in a network to detect HOIs and can provide a reference for modeling contextual information in the HOI problem.
- EfficientFAN: Deep Knowledge Transfer for Face Alignment5
A lightweight, efficient, and effective model for face alignment that achieves equivalent performance with fewer model parameters, lower computation cost, and higher speed is proposed and named EfficientFAN.
- Energy minimisation‐based multi‐class multi‐instance geometric primitives extraction from 3D point clouds5
A novel energy minimisation-based algorithm for multi-class multi-instance geometric primitives extraction from the 3D point cloud is proposed and can automatically determine the classes and numbers of geometricPrimitives in the 3d point cloud.
- 5
- Dense Attention Network for Facial Expression Recognition in the Wild5
This paper proposes an attention mechanism based module which can help the network focus on the emotion-related locations and produces two network structures named DenseCANet and DenseSANet by using the attention modules based on the backbone of DenseNet.
- 5
- 5
- Collaborative Q-Learning Based Routing Control in Unstructured P2P Networks5
The proposed collaborative Q-learning method learns the networks parameters such as processing capacity, number of connections, and number of resources in the peers, along with their state of congestion, to assure alternative routing path to balance the query loads among the peers under higher network churns.
- 5
- A New Interactive Segmentation Scheme Based on Fuzzy Affinity and Live-Wire5
The combination of the Live-Wire method with the region growing algorithm based on fuzzy affinity is reported, which has been used for CT and MR image segmentation and confirmed that this method is practical and accurate in the medical images segmentation.
- Semi-supervised Object Detection for Remote Sensing Images via Pseudo Labeling and Consistency Learning4
RS-PCL is proposed, which is a semi-supervised method applying Pseudo labeling and Consistency Learning to RSI, which uses the teacher-student training scheme to generate high-quality pseudo labels for self-training.
- 4
- Human Pose Estimation based on Attention Multi-resolution Network4
An attention-mechanism-based multi-resolution network is proposed, which adds an attention mechanism to the High-Resolution Network (HRNet) to enhance the feature representation of the network.
- An Automated Lung Nodule Segmentation Method Based On Nodule Detection Network and Region Growing4
An automated segmentation method is proposed for lung nodules in CT images that is evaluated on a public lung nodule dataset, and experimental results indicate the effectiveness and efficiency of the proposed method.
- 4
- Spatiotemporal text localization for videos4
A unified framework is proposed which consists of the sampling- and-recovery model (SaRM) and the divide-and-conquer model (DaCM), designed to efficiently localize the text in spatiotemporal domain simultaneously.
- 4
- Weakly-supervised scene parsing with multiple contextual cues4
A weakly-supervised scene parsing algorithm that semantically parses a collection of images with multi-label, which is guided by the top-down category models and bottom-up local patch contexts across images that closely related segments usually have similar labels.
- Nonlocal variational image segmentation models on graphs using the Split Bregman4
An image segmentation model in a variational nonlocal means framework based on a weighted graph that is superior in terms of both accuracy and efficient is proposed.
- Phrasal Paraphrase Based Question Reformulation for Archived Question Retrieval4
This paper presents a question reformulation scheme to enhance the question retrieval model by fully exploring the intelligence of paraphrase in phrase-level, which compensates for the existing paraphrasing research in a suitable granularity.
- 4
- A new deformable model using level sets for shape segmentation4
A new deformable model for shape segmentation is presented, which makes two modifications to the original level set implementation of deformable models to replace the edge detector by some area constraint, and the second modification utilizes weighted length constraint to regularize the curve under evolution.
- 3
- 3
- 3
- Gated Multi-Modal Edge Refinement Network for Light Field Salient Object Detection3
A gated multi-modal edge refinement network (GMERNet), which first obtains the preliminary position and structure information of the salient object and then gradually refines the object edge, to tackle the challenges of inaccurate edge prediction and effectively leverage the rich multi-modal light field information.
- 3
- SCU-Net: shape constraint U-Net for prostate segmentation in MR images3
An automatic segmentation model called the Shape Constraint U-Net (SCU-Net) for prostate MR images is proposed, which substantially improves prostate boundary region segmentation and can be used to improve the boundary voxel prediction performance of other segmentation networks.
- 3
- 3
- 3
- 3
- 3
- Convex‐relaxed active contour model based on localised kernel mapping3
This study presents a localised KM (LKM) method based on the fact that intensity inhomogeneity can be ignored in a local neighbourhood and first reconstructs the KM formulation of image segmentation in a neighbourhood of each pixel, and then such formulations can be integrated together to derive the LKM energy functional.
- Hybrid architecture for 3D visualization of ultrasonic data3
The experimental results indicate that the proposed hybrid architecture and methods are effective and efficient in processing and visualizing very large-volumetric datasets.
- Automatic Segmentation of Extraocular Muscle Using Level Sets Methods with Shape Prior3
This paper presents a fully automatic segmentation method for extraocular muscle based on the level set scheme while integrating the prior information of shape and pose while demonstrating the efficiency of the proposed segmentation procedure.
- Inscription extraction from Traditional Chinese Painting images3
This proposed algorithm analyzes the existing text extraction algorithms for natural image and combines the features of inscription in the TCP images to extract the inscription area in Traditional Chinese Painting (TCP) images.
- Image-based 3D models reconstruction3
This paper proposes a complete system that covers the steps from image capture to final visual surface reconstruction for objects in an integrated 3D model reconstruction system.
- A Geometric Active Contours Model for Multiple Objects Segmentation3
An automatic geometric active contours model which can extract multiple objects in an image without any manual assistance and completely eliminates the need of costly re-initialization procedure is presented.
- 3
- 2
- MotionFlow: Joint Motion Priors and Appearance Enhancement for High-Accuracy Optical Flow Estimation2
- 2
- 2
- GPDF-Net: geometric prior-guided stereo matching with disparity fusion refinement2
A geometric prior-guided stereo matching method with disparity fusion refinement, named GPDF-Net, is proposed, which achieves competitive performance with respect to both accuracy and robustness and produces better results than those of other methods when given blurred edges and textureless and occluded areas.
- 2
- 2
- 2
- 2
- Uncertainty-Aware Semi-Supervised Learning of 3D Face Rigging from Single Image2
A method to rig 3D faces via Action Units, viewpoint and light direction, from single input image by integrating a recent 3D reconstruction method with 2D AU-driven method in a semi-supervised fashion is presented.
- 2
- Saliency Detection for Compressive Sensing Measurements2
This work proposes a method of saliency detection using compressive sensing global measurement values, which uses the characteristics of the Hadamard matrix to reconstruct the original image with low resolution, and then uses the low resolution image forSaliency detection.
- 2
- 2
- Image Denoising Using Fractional-Order Non-Local TV Model2
An efficient numerical method based on the dual theory is introduced to solve the corresponding Euler-Lagrange equation and the relationship between the regularization parameter and fractional-order is analyzed, which provides an important basis for the design of optimal strategy.
- A Coastline Detection Method Based on Level Set2
This method is an improvement on the traditional level set algorithm by using the information of GSHHS to optimize the initialization procedure, and optimizes regional energy functional, and can achieve the rapid coastline detection.
- An Image Segmentation Method for Chinese Paintings by Combining Deformable Models with Graph Cuts2
A fast object extraction method that combines deformable models information with explicit edge information in a graph cuts optimization framework, and introduces the texture information by building TCP images' co-occurrence matrices.
- 2
- 1
- QueryGaussian: Scalable and Training-Free Open-Vocabulary 3D Instance Retrieval1
This work proposes QueryGaussian, a training-free framework for expeditious and scalable open-vocabulary 3D instance retrieval that enables expeditious instance retrieval on city-scale scenes containing tens of millions of Gaussians using consumer-grade hardware.
- 1
- 1
- 1
- 1
- 1
- UPFP-SG: A New Benchmark for Unilateral Peripheral Facial Paralysis Severity Grading1
This work proposes a new benchmark named UPFG-SG for Unilateral Peripheral Facial Paralysis Severity Grading, and proposes a new method trained on this dataset which integrates different facial features to rate the facial palsy severity of each facial nerve region.
- Deep Learning for Distinguishing Mucinous Breast Carcinoma From Fibroadenoma on Ultrasound1
The DL+ age-tree model based on US images and age showed excellent performance in the differentiation of MBC and FA and can effectively improve the performance of radiologists with different degrees of experience that may contribute to reducing the misdiagnosis of MBC.
- 1
- 1
- 1
- 1
- 1
- Dual Optical Path Based Adaptive Compressive Sensing Imaging System1
Experiments have proved that the method proposed can filter the noise in the sampling process of the CS system, and can significantly improve the quality of image reconstruction under a variety of algorithms.
- 1
- A Lightweight Gated Global Module for Global Context Modeling in Neural Networks1
A novel gated global module (GGM) that is lightweight and yet effective in terms of achieving better integration of global information in relation to feature representation and can be easily integrated with common CNN architectures and is training friendly.
- 1
- Extraction of Multi-class Multi-instance Geometric Primitives from Point Clouds Using Energy Minimization1
A novel algorithm for extracting multiple instances of multiple classes of geometric primitives from point clouds by using a new sampling strategy and an improved optimization technique to minimize the energy.
- Affine Collaborative Representation Based Classification for In-Air Handwritten Chinese Character Recognition1
This paper presents a study of using affine collaborative representation based classification (ACRC) for recognizing the in-air handwritten Chinese characters, which is collected by Leap Motion, one kind of 3-dimensional sensing device.
- Ringing Reduction in Image Restoration Using Cyclic Boundary1
A method to reduce ringing for image restoration based on the Richardson-Lucy algorithm and cyclic boundaries is proposed and experimental results show that the proposed algorithm achieves better ringing reduction than the other two algorithms.
- Local kernel mapping based piecewise constant model for medical image segmentation1
The proposed model has higher segmentation accuracy in the presence of intensity inhomogeneity and noise and energy minimization has been implemented in level set framework.
- Detecting Salient Objects via Spatial and Appearance Compactness Hypotheses1
This work proposes a novel compactness hypothesis which assumes salient regions are more compact than background spatially and in appearance and implements an effective object-level saliency detection method, which is demonstrated to be effective even in difficult cases.
- 1
- Image deconvolution using ℓ 1 sparse regularization1
This paper studies sparse regularization image deconvolution scheme over the space of measures, the natural extension of the ℓ1 norm of vectors to the setting of measures and gets a simple iterative algorithm.
- Aluminum CT Image Defect Detection Based on Segmentation and Feature Extraction1
This paper presents the approach to detecting defects follows a general image processing scheme based on three steps: segmentation, feature extractions, and classification, which is fast and accurate to defects detection in CT image, and the method has high robustness for illumination.
- A transmission scheme for opportunistic networks with social selfish nodes1
In the proposed scheme, a priority is given to a message according to the cooperation degree, a cache management scheme is applied, and a delivery probability is calculated to guide nodes to select the better nodes as its relay.
- A Sufficient Condition for Uniform Convergence of Stationary p-Subdivision Scheme1
According to the properties of supported set of the nonnegative mask, a sufficient condition for uniform convergence of stationary p-subdivision scheme is presented and this condition is proved with two propositions and spline function.
- Oriented Teacher: Exploiting Spatial Resolution Disparity to Enhance Semi-Supervised Aerial Object Detection–
This study introduces a novel framework, Oriented Teacher, which incorporates a Multi-resolution Teaching Strategy (MTS), and employs a combination of image transformations to adaptively adjust the spatial resolution of the unlabeled image, ensuring better alignment with labeled images and higher-quality pseudo labels generated by the teacher model.
- –
- –
- –
- IoU-PD: IoU-Aware Privileged Distillation for Visual Grounding with Multimodal Large Language Models–
Experiments on standard referring-expression grounding benchmarks show consistent region-level improvements over strong coordinate-generating baselines, demonstrating that ground-truth boxes can provide useful privileged guidance beyond serving as coordinate labels.
- –
- SCG-SSC: Semantic Scene Completion via Self-and-Cross Gated Fusion of Depth Maps and Semantic Priors–
- –
- –
- –
- –
- –
- Conditional flow matching with coordinate-aware patch training for efficient 3D medical image segmentation–
This work proposes a conditional Flow Matching (CFM) based framework for efficient 3D medical image segmentation that maintains competitive segmentation accuracy while achieving an inference time of only 1.14 seconds per volume, offering a clear efficiency advantage over existing generative segmentation methods.
- Frequency-Spatial Collaborative Matching for Cross-Domain One-Shot Medical Image Segmentation–
The Frequency-Spatial Collaborative Matching Network (FSCMNet) is proposed, which refines the coarse query mask via pixel-level supervised contrastive learning to improve feature discriminability and performs collaborative matching by jointly aligning support-query features in both the mid-frequency domain and the spatial domain.
- –
- –
- –
- –
- –
- Motion Reconstruction via Human Anatomy Diffusion from Sparse Tracking–
This supplementary material delves deeper into the architectural and experimental nuances that complement the main study, shedding light on hyperparameters’ effects, the components’ order and structure.
- –
- –
- –
- Towards Accurate and Smooth Human Motion Reconstruction from Sparse Tracking with State Space Model–
A novel conditional diffusion model, Pose Mamba Diffusion (PMDiff), which comprises a bidirectional Pose Mamba Denoiser to achieve effective sequence modeling and outperforms previous state-of-the-art methods in terms of both accuracy and smoothness.
- –
- –
- –
- –
- AdaptControl: Adaptive Human Motion Control and Generation via User Prompt and Spatial Trajectory Guidance–
AdaptControl is proposed, a motion generation model capable of precise spatial trajectory control while ensuring that the generated motion adheres to both text descriptions and spatial guidance, opening up new possibilities for intuitive and interactive human motion generation and facilitating the development of user-centered design and interactive applications.
- –
- –
- –
- –
- –
- Joint segmentation and registration for 4D lung CT images based on Markov random field–
A general simultaneous segmentation and registration framework based on MRF (Markov Random Field) is proposed, which utilizes the segmentation results of one frame as an atlas for the initialization step and a stochastic sampling approach is introduced for the computation of registration similarity measurement.
- –
- –
- Network Scheduling Model of Cloud Computing–
The results showed, compared with traditional intelligence Particle Swarm Optimization algorithms, the improved algorithm can preferably allocate the resources in cloud computing model, the effect of prediction model time is more close to actual time, and the optimal solution’s time of objective function value is shorten which meet the user's needs more.
- –
- Semi-supervised Remote Sensing Image Segmentation Using Dynamic Region Merging–
Comparisons with the state-of-the-art methods demonstrate that the proposed method can effectively investigate the spatial relationship among pixels and achieve better remote sensing image segmentation results.
- Variational nonlocal image segmentation using split-Bregman–
An image segmentation model in a variational nonlocal means framework that combines the nonlocal regularization and the local intensity fitting terms to demonstrate accuracy and display superiority.
- –
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.