Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks263 from public data
- Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition12,821
This work equips the networks with another pooling strategy, “spatial pyramid pooling”, to eliminate the above requirement, and develops a new network structure, called SPP-net, which can generate a fixed-length representation regardless of image size/scale.
- 5,457
- 3,149
- Channel Pruning for Accelerating Very Deep Neural Networks2,839
This paper proposes an iterative two-step algorithm to effectively prune each layer, by a LASSO regression based channel selection and least square reconstruction, and generalizes this algorithm to multi-layer and multi-branch cases.
- 2,617
- 1,450
- 1,330
- Deep ADMM-Net for compressive sensing MRI1,287
Experiments on MRI image reconstruction under different sampling ratios in k-space demonstrate that the proposed novel ADMM-Net algorithm significantly improves the baseline ADMM algorithm and achieves high reconstruction accuracies with fast computational speed.
- Lazy snapping1,276
Usability studies indicate that Lazy Snapping provides a better user experience and produces better segmentation results than the state-of-the-art interactive image cutout tool, Magnetic Lasso in Adobe Photoshop.
- 1,255
- Learning a convolutional neural network for non-uniform motion blur removal930
A deep learning approach to predicting the probabilistic distribution of motion blur at the patch level using a convolutional neural network (CNN) is proposed and the candidate set of motion kernels predicted by the CNN are extended using carefully designed image rotations.
- 908
- Geodesic Saliency Using Background Priors885
Evaluation on two databases validates that geodesic saliency achieves superior results and outperforms previous approaches by a large margin, in both accuracy and speed (2 ms per image), illustrating that appropriate prior exploitation is helpful for the ill-posed saliency detection problem.
- 883
- Accelerating Very Deep Convolutional Networks for Classification and Detection860
This paper aims to accelerate the test-time computation of convolutional neural networks, especially very deep CNNs, and develops an effective solution to the resulting nonlinear optimization problem without the need of stochastic gradient descent (SGD).
- Image deblurring with blurred/noisy image pairs764
This paper shows in this paper how to produce a high quality image that cannot be obtained by simply denoising the noisy image, or deblurring the blurred image alone, by combining information extracted from both blurred and noisy images.
- 699
- Image completion with structure propagation671
This paper introduces a novel approach to image completion in which the user manually specifies important missing structure information by extending a few curves or line segments from the known to the unknown regions by adopting the Belief Propagation algorithm to find the optimal patches.
- 546
- 524
- 480
- Object Detection Networks on Convolutional Feature Maps454
It is shown by experiments that despite the effective ResNets and Faster R-CNN systems, the design of NoCs is an essential element for the 1st-place winning entries in ImageNet and MS COCO challenges 2015.
- 437
- Bayesian Face Revisited: A Joint Formulation436
This paper revisits the classical Bayesian face recognition method by Baback Moghaddam et al. and proposes a new joint formulation that leads to an EM-like model learning at the training time and an efficient, closed-formed computation at the test time.
- 415
- Joint Cascade Face Detection and Alignment412
The key idea is to combine face alignment with detection, observing that aligned face shapes provide better features for face classification and learns the two tasks jointly in the same cascade framework, by exploiting recent advances in face alignment.
- Bundled camera paths for video stabilization410
A novel video stabilization method which models camera motion with a bundle of (multiple) camera paths based on a mesh-based, spatially-variant motion representation and an adaptive, space-time path optimization and introduces the 'as-similar-as-possible' idea to make motion estimation more robust.
- 401
- Video object cut and paste356
This paper presents a system for cutting a moving object out from a video clip using a new 3D graph cut based segmentation approach on the spatial-temporal video volume and provides brush tools for the user to control the object boundary precisely wherever needed.
- 341
- 333
- Focal Sparse Convolutional Networks for 3D Object Detection326
This paper introduces two new modules to enhance the capability of Sparse CNNs, both are based on making feature sparsity learnable with position-wise importance prediction and shows that spatially learnable sparsity in sparse convolution is essential for sophisticated 3D object detection.
- Proximal Dehaze-Net: A Prior Learning-Based Deep Network for Single Image Dehazing315
A novel deep learning approach for single image dehazing by learning dark channel and transmission priors and incorporating haze-related prior learning into deep network is proposed.
- 310
- 304
- 296
- Global Intrinsic Symmetries of Shapes288
An algorithm is devised which detects and computes the isometric mappings from the shape onto itself and is both computationally efficient and robust with respect to small non‐isometric deformations, even if they include topological changes.
- 272
- Progressive inter-scale and intra-scale non-blind image deconvolution271
This paper presents a progressive inter-scale and intra-scale non-blind image deconvolution approach that significantly reduces ringing and is built on a novel edge-preserving deconVolution algorithm called bilateral Richardson-Lucy (BRL) which uses a large spatial support to handle large blur.
- Background Cut254
Using background cut, a high quality and real-time foreground layer extraction algorithm that combines background subtraction, color and contrast cues to extract a foreground layer accurately and efficiently, the layer extraction errors caused by background clutter can be substantially reduced.
- Discrete laplace operator on meshed surfaces246
This paper proposes the first algorithm for approximating the Laplace operator of a surface from a mesh with point-wise convergence guarantees applicable to arbitrary meshed surfaces and shows that for a sufficiently fine mesh over an arbitrary surface, the mesh Laplacian is close to the La place-Beltrami operator on the surface at every point of the surface.
- 240
- 222
- 211
- Statistics of Patch Offsets for Image Completion211
This paper observation that if the authors match similar patches in the image and obtain their offsets (relative positions), the statistics of these offsets are sparsely distributed means that a few dominant offsets provide reliable information for completing the image.
- 210
- Cat Head Detection - How to Effectively Exploit Shape and Texture Features206
This paper proposes a two step approach for the cat head detection and proposes a set of novel features based on oriented gradients, which outperforms existing leading features, e.
- 205
- Automatic Exposure Correction of Consumer Photographs189
This paper will automate the interactive correction technique by estimating the image specific S-shaped non-linear tone curve that best fits the input image by creating a new Zone-based region-level optimal exposure evaluation, which would consider both the visibility of individual regions and relative contrast between regions.
- 189
- 182
- Unpaired Brain MR-to-CT Synthesis Using a Structure-Constrained CycleGAN180
A structure-constrained cycleGAN is proposed for brain MR-to-CT synthesis using unpaired data that defines an extra structure-consistency loss based on the modality independent neighborhood descriptor to constrain structural consistency.
- 178
- Face Alignment Via Component-Based Discriminative Search173
This paper proposes a component-based discriminative approach for face alignment without requiring initialization that searches the face shape in a large range at the component level by a discriminatives search algorithm.
- 172
- 172
- 166
- MeshFlow: Minimum Latency Online Video Stabilization160
The quantitative and qualitative evaluations show that the proposed technique for online video stabilization with only one frame latency using a novel MeshFlow motion model can produce comparable results with the state-of-the-art off-line methods.
- 158
- 157
- 149
- Finding Matches in a Haystack: A Max-Pooling Strategy for Graph Matching in the Presence of Outliers146
- Fast image deconvolution using closed-form thresholding formulas of regularization145
Based on the closed-form formulas for L"q(q=12,23) regularization, a fast algorithm to solve the image deconvolution problem using half-quadratic splitting method is proposed and extensive experiments demonstrate that the algorithm has a significant acceleration over Krishnan et al.'s algorithm.
- 143
- Topological methods for exploring low-density states in biomolecular folding pathways143
A computational approach to explore the relatively low populated transition or intermediate states in biomolecular folding pathways, based on a topological data analysis tool, MAPPER, with simulation data from large-scale distributed computing, inspired by classical Morse theory in mathematics.
- 140
- 133
- 132
- 131
- 130
- Rich Image Captioning in the Wild130
An image caption system that addresses new challenges of automatically describing images in the wild by developing a deep vision model that detects a broad range of visual concepts, an entity recognition model that identifies celebrities and landmarks, and a confidence model for the caption output.
- 123
- Voxel Field Fusion for 3D Object Detection121
This work presents a conceptually simple yet effective framework for cross-modality 3D object detection, named voxel field fusion, and develops mixed augmentor to align feature-variant transformations, which bridges the modality gap in data augmentation.
- Fast burst images denoising120
A fast denoising method that produces a clean image from a burst of noisy images by introducing a lightweight camera motion representation called homography flow and a mechanism of selecting consistent pixels for temporal fusion to handle scene motion during the capture.
- Computing geometry-aware handle and tunnel loops in 3D models117
The algorithm is a novel application of the concepts from topological persistence introduced recently in computational topology by computing topologically correct loops that are also geometrically relevant.
- 116
- 106
- Rectangling panoramic images via warping103
This paper presents a content-aware warping algorithm that generates rectangular images from stitched panoramic images, and demonstrates that the results are often visually plausible, and the introduced distortion is often unnoticeable.
- Structural Insight into RNA Hairpin Folding Intermediates101
This work studies the folding of a small tetraloop hairpin using a serial version of replica exchange molecular dynamics on a distributed computing environment and finds that folding is not simply the reverse of high-temperature unfolding.
- Well Begun Is Half Done: Generating High-Quality Seeds for Automatic Image Dataset Construction from Web100
This work proposes a density score based on rank-order distance to identify positive seed images and guarantees the selected seeds are as numerous and accurate as possible through adaptive thresholding.
- 99
- 98
- 95
- NTIRE 2022 Challenge on Efficient Super-Resolution: Methods and Results94
The NTIRE 2022 challenge was to super-resolve an input image with a magnification factor of ×4 based on pairs of low and corresponding high resolution images and the aim was to design a network for single image super-resolution that achieved improvement of efficiency measured according to several metrics.
- SkyFinder87
This paper presents SkyFinder, an interactive search system that computes all sky attributes offline, then provides an interactive online search engine and builds a sky graph based on the sky attributes, so that the user can smoothly explore and find a path within the space of skies.
- 84
- 83
- 82
- Poisson image fusion based on Markov random field fusion model81
A gradient domain image fusion framework based on the Markov Random Field (MRF) fusion model that is able to better fuse the multi-sensor images and produces high-quality fusion results compared with the other state-of-the-art methods.
- Interactive normal reconstruction from a single image77
An interactive system for reconstructing surface normals from a single image by introducing a novel shape-from-shading algorithm (SfS) that produces faithful normal reconstruction for local image region, but it fails to faithfully recover the overall global structure.
- 77
- Bayesian Correction of Image Intensity with Spatial Consideration75
A novel approach to recover a high-quality image by exploiting the tradeoff between exposure time and motion blur, which considers color statistics and spatial constraints simultaneously, by using only two defective input images.
- 74
- Point Integral Method for Solving Poisson-Type Equations on Manifolds from Point Clouds with Convergence Guarantees73
The key idea is to derive the integral equations which approximates the Poisson-type equations and contains no derivatives but only the values of the unknown function, which makes the integral equation easy to be approximated from point cloud.
- 73
- Graph Cuts for Supervised Binary Coding71
This paper forms supervised binary coding as a single optimization problem that involves both the encoding functions and the binary label assignment, and applies the graph cuts algorithm to address the discrete optimization problem involved, with no continuous relaxation.
- 71
- Hierarchical Nyström methods for constructing Markov state models for conformational dynamics69
This paper proposes a formalism based on an algebraic principle for matrix approximation, i.e., the Nyström method, to deal with poorly sampled microstates and demonstrates the ability of this algorithm to discover the major metastable states on two model systems, the alanine dipeptide and trpzip2 peptide.
- 68
- 66
- 65
- CONSTRUCTING MULTI-RESOLUTION MARKOV STATE MODELS (MSMS) TO ELUCIDATE RNA HAIRPIN FOLDING MECHANISMS59
- 57
- 57
- Approximating loops in a shortest homology basis from point data57
An algorithm to compute a set of loops from a point data that presumably sample a smooth manifold M ⊂ Rd to approximate a shortest basis of the one dimensional homology group H1(M) over coefficients in finite field Z2.
- 56
- 53
- 52
- End-to-end Interpretable Learning of Non-blind Image Deblurring51
This work proposes to precondition the Richardson solver using approximate inverse filters of the (known) blur and natural image prior kernels to allow extremely efficient parameter sharing across the image, and leads to significant gains in accuracy and/or speed compared to classical FFT and conjugate-gradient methods.
- Deep Positional and Relational Feature Learning for Rotation-Invariant Point Cloud Analysis48
A novel deep network for point clouds analysis is proposed by incorporating positional information of points as inputs while yielding rotation-invariance, and experiments show state-of-the-art classification and segmentation performances on benchmark datasets.
- Normal and Feature Approximations from Noisy Point Clouds48
This work considers the problem of approximating normal and feature sizes of a surface from point cloud data that may be noisy and provides new analyses that lead to new algorithms for practical and reliablenormal and feature approximations.
- NTIRE 2022 Challenge on High Dynamic Range Imaging: Methods and Results45
This paper reviews the challenge on constrained high dynamic range (HDR) imaging that was part of the New Trends in Image Restoration and Enhancement (NTIRE) workshop, held in conjunction with CVPR 2022, and the competition set-up, datasets, proposed methods and their results.
- 45
- Dynamic Grained Encoder for Vision Transformers44
A Dynamic Grained Encoder for vision transformers is proposed, which can adaptively assign a suitable number of queries to each spatial region and achieves a fine-grained representation in discriminative regions while keeping high efficiency.
- 43
- NTIRE 2021 Challenge on Burst Super-Resolution: Methods and Results43
The NTIRE2021 challenge on burst super-resolution was to generate a clean RGB image with 4 times higher resolution given a RAW noisy burst as input, and the top-performing methods set a new state-of-the-art for the burstsuper-resolution task.
- 41
- 3D Point Cloud Recognition Based on a Multi-View Convolutional Neural Network41
This paper designs an effective recognition model based on a multi-view convolutional neural network that can effectively solve the recognition problem of lidar point clouds and provide vital practical value.
- 41
- 40
- A fast parallel clustering algorithm for molecular simulation trajectories40
A GPU‐powered parallel k‐centers algorithm to perform clustering on the conformations of molecular dynamics (MD) simulations that is up to two orders of magnitude faster than the CPU implementation and finds the triangle inequality to be less effective in higher dimensions.
- 40
- 39
- 39
- Fast Camera Image Denoising on Mobile GPUs with Deep Learning, Mobile AI 2021 Challenge: Report39
The first Mobile AI challenge is introduced, where the target is to develop an end-to-end deep learning-based image denoising solution that can demonstrate high efficiency on smartphone GPUs.
- 38
- Surface Reconstruction Based on the Modified Gauss Formula38
A surface reconstruction method that has excellent performance despite nonuniformly distributed, noisy, and sparse data is introduced and can be parallelized with small overhead and shows compelling performance in a GPU version by implementing this direct and simple approach.
- Gromov-Hausdorff Approximation of Filament Structure Using Reeb-type Graph38
It is proved that filamentary structures that can be seen as topological metric graphs can be approximated, with respect to the Gromov-Hausdorff distance by well-chosen Reeb graphs (and some of their variants) and an efficient and easy to implement algorithm to compute such approximations is provided.
- 38
- 37
- 37
- Neural multi-atlas label fusion: Application to cardiac MR images35
The proposed novel multi‐atlas segmentation method, dubbed deep fusion net (DFN), is a deep architecture that integrates a feature extraction subnet and a non‐local patch‐based label fusion (NL‐PLF) subnet in a single network.
- 35
- 34
- A Unified Hyper-GAN Model for Unpaired Multi-contrast MR Image Translation34
A unified Hyper-GAN model for effectively and efficiently translating between different contrast pairs, which consists of a pair of hyper-encoder and hyper-decoder to first map from the source contrast to a common feature space, and then further map to the target contrast image.
- 32
- 31
- Approximating cycles in a shortest basis of the first homology group from point data31
An algorithm is presented to compute a set of cycles from a point data that presumably sample a smooth manifold that approximate a shortest basis of the first homology group over coefficients in the finite field.
- 31
- Learning Dictionary of Discriminative Part Detectors for Image Categorization and Cosegmentation30
A stochastic version of a proximal algorithm to solve the corresponding optimization problem of category-specific part detectors for image categorization and cosegmentation.
- 30
- A New Barrier Certificate for Safety Verification of Hybrid Systems30
A new barrier certificate condition, called Exponential Condition, is proposed, for the safety verification of semialgebraic hybrid systems, which has a lower conservativeness than the existing convex conditions and meanwhile it possesses the convexity.
- 28
- 27
- Effective Clipart Image Vectorization through Direct Optimization of Bezigons27
This paper describes a novel method for vectorizing clipart images that directly optimize the bezigons rather than using other intermediate representations and outperforms both the current state-of-the-art method and commonly used commercial software in terms of bezigon quality.
- HyperAdam: A Learnable Task-Adaptive Adam for Network Training26
A new optimizer, dubbed as HyperAdam, is proposed that combines the idea of "learning to optimize" and traditional Adam optimizer and is justified to be state-of-the-art for various network training, such as multilayer perceptron, CNN and LSTM.
- 26
- Model-Driven Deep Attention Network for Ultra-fast Compressive Sensing MRI Guided by Cross-contrast MR Image25
A model-driven deep attention network is proposed, dubbed as MD-DAN, to reconstruct highly under-sampled long-time sequence MR image with the guidance of a certain short-time sequences MR image to accelerate MRI.
- Gromov–Hausdorff Approximation of Filamentary Structures Using Reeb-Type Graphs25
It is proved that filamentary structures that can be seen as topological metric graphs can be approximated with respect to the Gromov–Hausdorff distance by well-chosen Reeb graphs and an efficient and easy-to-implement algorithm is provided to compute such approximations in almost linear time.
- 23
- SHREC'10 Track: Feature Detection and Description23
The present paper is a report of the 3D Shape Retrieval Contest 2010 (SHREC'10) feature detection and description benchmark results.
- Learning Canonical View Representation for 3D Shape Recognition with Arbitrary Views22
A canonical view representation is proposed to tackle the challenge of recognizing 3D shapes from arbitrary views, i.e., arbitrary numbers and positions of viewpoints, and achieves competitive results under the fixed viewpoint settings and significantly outperforms the applicable methods under the arbitrary view setting.
- Convergence of the point integral method for Laplace–Beltrami equation on point cloud22
The convergence of PIM for Poisson equation with Neumann boundary condition on submanifolds that are isometrically embedded in Euclidean spaces is analyzed.
- Fuzzy Geodesics and Consistent Sparse Correspondences For: eformable Shapes21
A new concept called fuzzy geodesics is presented and it is shown that fuzzy geodeics are stable with respect to the Gromov‐Hausdorff distance and a new object called the intersection configuration for a set of points on a shape is proposed.
- 18
- Harmonic Extension on The Point Cloud18
The harmonic extension problem is considered, which is widely used in many applications of machine learning, and is formulated as solving a Laplace--Beltrami equation.
- 18
- 17
- Method for monitoring the cotton plant vigor based on the WSN technology17
A new method for monitoring the plant vigor is proposed, and the fusion data based on fuzzy comprehensive evaluation improves the accuracy of judging the vigor.
- Surface reconstruction from unorganized points with l0 gradient minimization16
A novel efficient and fast method by using l0 gradient minimization, which can directly measure the sparsity of a solution and produce sharper surfaces is proposed, which is particularly effective for sharpening major edges and removing noise.
- 16
- 15
- 15
- A loop closure improvement method of Gmapping for low cost and resolution laser scanner15
A Kalman filter based algorithm is proposed to improve the loop closure correction performance of Gmapping using low cost and resolution laser scanners (e.g. RPLidar laser scanner), which can positively promote the application of laser scanner in normal life.
- 15
- A Model-Driven Deep Dehazing Approach by Learning Deep Priors14
This paper designs a deep dehazing neural network, dubbed as proximal dehaze-net, by learning the proximal operators for haze-related image priors using CNNs, which incorporates physical model constraints of hazes and haze- related prior learning into a novel deep architecture.
- 14
- 13
- A Prior Learning Network for Joint Image and Sensitivity Estimation in Parallel MR Imaging13
A novel deep network is proposed, dubbed as Blind-PMRI-Net, to simultaneously reconstruct the MR image and sensitivity maps in a blind setting for parallel imaging, which naturally combines the physical constraint of parallel imaging and prior learning in a single deep architecture.
- 13
- 13
- Scale selection for anisotropic diffusion filter by Markov random field model13
This paper designs a Markov random field (MRF) scale selection model, which selects scales for image segments, then the denoised image is the composition of segments at their optimal scales in the scale space.
- 12
- Abnormality detection in retinal image by individualized background learning12
This paper tries to identify diverse abnormalities from a retinal test image by finely learning its individualized retinal background (IRB) on which retinal lesions superimpose, using a multi-scale sparse coding based learning (MSSCL) algorithm and a repeated learning strategy.
- Model Meets Deep Learning in Image Inverse Problems12
This paper will review a new trend of methods for image inverse problem that combines the imaging/degradation model with deep learning approach, typically designed by unrolling some optimization algorithms or statistical inference algorithms into deep neural networks.
- Diverse lesion detection from retinal images by subspace learning over normal samples12
Inspired by the doctors’ visual diagnostic mode, this paper proposes a novel computational framework to detect various types of lesions from retinal images, in which many healthy fundus images are collected to act as ”doctors’ detection experience" and local visual properties of lesions are used to distinguish true positives from false positives.
- Convergence of the Point Integral Method for the Poisson Equation on Manifolds II: the Dirichlet Boundary12
This paper proves the convergence of the point integral method for solving the Poisson equation with the Dirichlet boundary condition.
- Training Networks in Null Space of Feature Covariance With Self-Supervision for Incremental Learning11
- LdCT-Net: low-dose CT image reconstruction strategy driven by a deep dual network11
This work presents a low-dose CT image reconstruction strategy driven by a deep dual network (LdCT-Net) to yield high-quality CT images by incorporating both projection information and image information simultaneously.
- 11
- Parallel adaptive large neighborhood search based on spark to solve VRPTW10
Aiming at the multi-objective vehicle path planning problem with time windows (VRPTW), a Spark-based parallel Adaptive Large Neighborhood Search algorithm (Spark-ALNS) is proposed to solve it and can find better solutions and find new optimal solutions for 31 algorithms, which outperforms other evolutionary algorithms.
- A real-time multi-constraints obstacle avoidance method using LiDAR10
A novel real-time multi-constraints obstacle avoidance method using LiDAR, which is able to find the sub-goal defined by a multi- Constraints function within the explored region and plan a corresponding optimal trajectory at each time step iteratively, so that the robot approaches the goal over time.
- 9
- 9
- 9
- A tensor-based nonlocal total variation model for multi-channel image recovery9
Extensive experimental results demonstrate that the proposed regularizer is systematically superior over other competing local and nonlocal regularization approaches, both quantitatively and visually.
- Path Plan of Unmanned Underwater Vehicle Using Particle Swarm Optimization9
Results indicates that particle swarm optimization based on swarm intelligence can insure optimality most out of limited computation time, relative to traditional method using graph theory and dynamic programming.
- Computing handle and tunnel loops with knot linking9
It is shown that a closed surface of genus g always has g handle and g tunnel loops induced by the embedding and this method is provided with applications to feature detection and topology simplification to show the effectiveness of the method.
- 8
- A Fourier-theoretic approach for inferring symmetries8
Using the Fourier-theoretic approach, an efficient marginal-based search strategy is constructed, which can recover the symmetry group G effectively and can fully determine the symmetries of various geometric objects.
- 7
- 7
- A new approach to solve inverse problems: Combination of model-based solving and example-based learning7
A new approach for solving the inverse problems which combines model solving and example-based learning is proposed in this work and applications demonstrate that the new method is not only feasible and efficient, but also has high generalization.
- 6
- Improved Baselines with Synchronized Encoding for Universal Medical Image Segmentation5
Sync SAM employs a synchronized dual-branch encoder that integrates convolution and Transformer features in a synchronized manner to enhance medical image encoding, and a multi-scale dual-branch decoder to preserve image details to optimize segmentation performance under standard inference settings.
- BaryIR: Learning Multi-Source Unified Representation in Continuous Barycenter Space for Generalizable All-in-One Image Restoration5
A multi-source representation learning framework BaryIR, which decomposes the latent space of multi-source degraded images into a continuous barycenter space for unified feature encoding and source-specific subspaces for specific semantic encoding, and exhibits superior generalization ability to real-world data and unseen degradations.
- 5
- 5
- 5
- Adversarial data splitting for domain generalization5
A novel domain generalization approach that enforces the learned model to be able to generalize well over the train/val subset splitting of the training dataset, formulated as a min-max optimization problem inspired by the meta-learning approach.
- 5
- 4
- Learning Polynomial-Based Separable Convolution for 3D Point Cloud Analysis4
This paper proposes a novel point convolution (PSConv) using separable weights learned with polynomials for 3D point cloud analysis, and proposes a hierarchical network (PSNet) defined on the point cloud, which achieves state-of-the-art accuracies for shape classification, as well as competitive results for shape segmentation compared with previous methods.
- Unlabeled data driven cost-sensitive inverse projection sparse representation-based classification with 1/2 regularization3
An improved SRC model is constructed to achieve effective and stable representation and recognition results and the cost-sensitive strategy is integrated into the model’s classification criteria to improve the imbalance of class distribution adaptively.
- Gaussian Processes in Polar Coordinates for Mobile Robot Using SE(2)-3D Constraints3
The Gaussian Process is developed to address insufficient scan data for low-precision 2D lidar by enriching the lidar measurements at interest or specific bearing regions and a new method, based on the graph optimization framework, to solve the non-SE(2) perturbations is proposed, namely SE2-3D constraint.
- Spherical Motion Dynamics: Learning Dynamics of Normalized Neural Network using SGD and Weight Decay3
- Extensive Comparison of Visual Features for Person Re-identification3
An extensive comparison for different kinds of visual features including hand-craft features and Convolutional Neural Networks features finds Colorhistogram and ScalableColor features are most robust to occlusion on CASIA, while GoogleNet and VG-GNet features have good robustness as well.
- 3
- 3
- 2
- Robust Offline Imitation Learning Through State-level Trajectory Stitching2
A state-based search framework that stitches state-action pairs from imperfect demonstrations, generating more diverse and informative training trajectories and a novel approach to enhance policy learning from mixed-quality offline datasets by leveraging task-relevant trajectory fragments and rich environmental dynamics.
- In vitro Isolation and Culturing of Mouse Primary Chondrocytes for Cartilage Biology2
This protocol provides a detailed description of the in vitro isolation and culturing methods for primary chondrocytes derived from mice, and shows the morphological characteristics and adhesion status of cells cultured for 1, 2, and 5 days after isolation.
- GMapLatent: Geometric Mapping in Latent Space2
This work innovate a cross-domain alignment and generation model that introduces a canonical latent space representation based on geometric mapping to align the cross-domain latent spaces in a rigorous and precise manner, thus avoiding mode collapse and mixture in the encoder-decoder generation architectures.
- Re-Visible Dual-Domain Self-Supervised Deep Unfolding Network for MRI Reconstruction2
A novel re-visible dual-domain self-supervised deep unfolding network is proposed to address issues when only under-sampled datasets are available, and outperforms state-of-the-art approaches in terms of reconstruction performance and generalization capability.
- A distribution independence based method for 3D face shape decomposition2
Both qualitative and quantitative experimental results show that the proposed approach can achieve state-of-the-art results in 3D face shape decomposition, and introduces a binary discriminator to enforce the latent representations of identity and expression to be distribution independent by adversarial learning.
- 2
- Learning Spectral Transform Network on 3D Surface for Non-rigid Shape Analysis2
This work proposes a novel spectral transform network on 3D surface to learn shape descriptors that achieved the highest accuracies on SHREC’14, 15 datasets as well as the “range” subset of SHREC'17 dataset.
- Study of Keeping Consistency of Chinese Corpus of Complete Parsing2
This paper analyzes the corpus which was amended by people and summarizes the reason leading to the annotation inconsistency and proposes strategies to eliminate inconsistencies and explored disambiguation type from various hierarchies.
- 2
- 2
- 2
- CSG-YOLO: a model for detecting minor and irregular steel surface defects based on deformable convolution and cross-layer fusion1
An improved YOLOv8 model–CSG-YOLO is proposed, which integrates deformable convolution, a Simplified Cross-stage Partial Spatial Pyramid Pooling Fusion (SimCSPSPPF) module, and a Global Feature Pyramid Network (GFPN), which aims to efficiently extract robust features of minor and irregular defects while meeting the stringent demands of industrial environments for real-time high-performance computing.
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- An Edge Preserving Regularization Model for Image Restoration Based on Hopfield Neural Network1
An edge preserving regularization model for image restoration that can significantly better preserve the edges of image compared with the commonly used Laplacian regularization (with constant and adaptive coefficient).
- 1
- –
- Point-LAG: local adaptive geometry-aware networks for LiDAR point cloud semantic segmentation–
A locally adaptive geometry-aware network that interactsively captures geometric and semantic features within the point cloud through a dual-path architecture to establish precise spatial relationships and achieve geometry-aware perception of local shapes is proposed.
- –
- –
- –
- –
- –
- –
- –
- –
- An adaptive censoring CFAR detector in non-homogeneous environments–
An adaptive censoring CFAR detector (AC-CFAR) is proposed, which first calculates the position of the transition and then identifies whether the subreference window starting from the transition is in a homogeneous environment.
- –
- –
- –
- Fast Lidar Inertial Odometry and Mapping for Mobile Robot SE(2) Navigation–
F-LIOM is shown to be cost-effective in terms of model complexity and computation efficiency for robot SE(2) navigation, as the motions in other degrees of freedom in 3D, including roll, pitch, and z, are considered to be noise terms that corrupt the pose estimation.
- –
- –
- –
- –
- –
- A Gradient-based Architecture HyperParameter Optimization Approach–
This work proposes a gradient-based approach to optimize these parameters in an efficient and unified manner, based on the observation that these parameters are consecutive and network performance changes continuously with them.
- Neural Embedding of an Iterative Deconvolution Algorithm for Motion Blur Estimation and Removal–
This paper introduces a new learning-based approach to motion blur removal that casts non-blind, non-uniform image deblurring as a least-squares problem regularized by natural image priors in the form of sparsity constraints.
- Image Fusion Algorithm Based on Deng Correlation and Image Saliency in Curvelet Framework–
The fusion framework is built on the second discrete curvelet transform to make full use of its multi-scale and multi-direction representability and subjective and objective evaluation results from experimental simulation show the efficacy of the proposed fusion algorithm.
- –
- –
- The Analysis on Materials Testing Network of Ecology Environment–
In order to accelerate the speed of the material testing, and ensure the quality and the accuracy of test results, characteristics and user's requirements must be selected and improved.
- Object Detection Based on Visual Selective Attention Mechanism–
The results prove that this method can avoid the leakage detection to some extent, which makes attention regional more close to the actuality of biological visual system.
- Chinese Chunking and Consistency Checking Using Rule-Based Method–
This paper extracts a small scale chunking rule set for chunking first, then defines more rules to check and correct the inconsistency phenomena and adopts man-machine interaction method to solve some special language phenomena.
- PSDV: A Neighbor Selection Policy in Unstructured P2P Systems–
The simulation experiment shows that the PSDV strategy can remarkably cut down the delay between peer and it's neighbors, lower the response time, and proves that the validity and rationality of the ideas are valid.
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.