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
- Azriel RosenfeldSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works337 from public data
- 947
- Shape similarity measure based on correspondence of visual parts532
This work applied a cognitively motivated similarity measure to shape matching of object contours in various image databases and compared it to well-known approaches in the literature, justifying that the shape matching procedure gives an intuitive shape correspondence and is stable with respect to noise distortions.
- 454
- 425
- Lednet: A Lightweight Encoder-Decoder Network for Real-Time Semantic Segmentation411
This paper presents LEDNet, which employs an asymmetric encoder-decoder architecture for the task of real-time semantic segmentation, which adopts a ResNet as backbone network to greatly reduce computation cost while maintaining higher segmentation accuracy.
- Convexity Rule for Shape Decomposition Based on Discrete Contour Evolution409
A novel rule is obtained, called the hierarchical convexity rule, which states that visual parts are enclosed by maximal convex (with respect to the object) boundary arcs at different stages of the contour evolution.
- 377
- Outlier Detection with Kernel Density Functions328
A novel unsupervised algorithm for outlier detection with a solid statistical foundation is proposed, modifying a nonparametric density estimate with a variable kernel to yield a robust local density estimation.
- 307
- GIFT: A Real-Time and Scalable 3D Shape Search Engine298
The proposed 3D shape search engine, which combines GPU acceleration and Inverted File (Twice), is named as GIFT, which outperforms the state-of-the-art methods significantly in retrieval accuracy on various shape benchmarks and competitions.
- 219
- 204
- Shape matching and classification using height functions193
A novel shape descriptor for matching and recognizing 2D object silhouettes that is not only invariant to geometric transformations such as translation, rotation and scaling but also insensitive to nonlinear deformations due to noise and occlusion is proposed.
- DeepMitosis: Mitosis detection via deep detection, verification and segmentation networks192
An accurate method for detecting the mitotic cells from histopathological slides using a novel multi‐stage deep learning framework and can achieve the highest F‐score on the MITOSIS dataset from ICPR 2012 grand challenge merely using the deep detection network.
- Bag of contour fragments for robust shape classification178
A new shape representation called Bag of Contour Fragments (BCF) inspired by classical Bag of Words (BoW) model is developed, which achieves the state-of-the-art performance on several well-known shape benchmarks, and can be applied to real image classification problem.
- Weakly supervised mitosis detection in breast histopathology images using concentric loss168
An automatic and accurate system for detecting mitosis in histopathology images using a deep segmentation network to produce segmentation map and a novel concentric loss function is proposed to train the semantic segmentsation network on weakly supervised mitosis data.
- 165
- 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.
- 153
- Multi-scale deep context convolutional neural networks for semantic segmentation132
This paper designs a multi-scale deep context convolutional network (MDCCNet), which combines the feature maps from different levels of network in a holistic manner for semantic segmentation and demonstrates that the method outperforms or is comparable to state-of-the-art methods on PASCAL VOC 2012 and SIFTFlow semantic segmentations datasets.
- AGLNet: Towards real-time semantic segmentation of self-driving images via attention-guided lightweight network129
An attention-guided lightweight network, namely AGLNet, which employs an encoder–decoder architecture for real-time semantic segmentation, and achieves state-of-the-art results in terms of speed and accuracy on three self-driving datasets.
- Well-Composed Sets127
It is proved that the skeletons obtained are "one point thick" and the formal definition of this concept is given, and it is shown that these skeletons have a graph structure and what this means.
- Balancing Deformability and Discriminability for Shape Matching121
This work presents the aspect shape context, which extends shape context-based descriptors and adaptively selects the "best" aspect weight for shape comparison, and shows that this framework naturally unifies existing important shape and image representations by adjusting the aspect weight and the embedding.
- Application of planar shape comparison to object retrieval in image databases116
Experimental results show that the shape matching procedure gives an intuitive shape correspondence and is stable with respect to noise distortions.
- 111
- 108
- Outlier Detection with Globally Optimal Exemplar-Based GMM107
A provably globally optimal Expectation Maximization (EM) algorithm is applied to flt a Gaussian Mixture Model (GMM) to a given data set, and the proposed outlier factor is based on global properties of the data set.
- 105
- 105
- Skeleton growing and pruning with bending potential ratio101
A novel significance measure for skeleton pruning, called bending potential ratio (BPR), in which the decision regarding whether a skeletal branch should be pruned or not is based on the context of the boundary segment that corresponds to the branch, is proposed.
- Schematizing Maps: Simplification of Geographic Shape by Discrete Curve Evolution101
An algorithm is presented, which on the one hand simplifies spatial data up to a degree of abstraction intended by the user; and on the other hand does not violate local spatial ordering between cartographic entities, since local arrangement of entities is assumed to be an important spatial knowledge characteristic.
- 96
- Shape guided contour grouping with particle filters94
A novel framework for contour based object detection and recognition is proposed, which is formulated as a joint contour fragment grouping and labeling problem, and shows that this framework yields not only accurate object detections but also localizations in real cluttered images.
- 94
- Contextual ensemble network for semantic segmentation91
A novel encoder-decoder architecture, called contextual ensemble network (CENet), for semantic segmentation, where the contextual cues are aggregated via densely usampling the convolutional features of deep layer to the shallow deconvolutional layers.
- 91
- 91
- 3D well-composed pictures91
There is only one connectedness relation in a well-composed picture, since 6, 18, and 26-connectedness are equivalent, which implies that many algorithms used in computer vision and computer graphics and their descriptions can be simpler, and the algorithms can be faster.
- Properties and Performance of Shape Similarity Measures90
This paper gives an overview of shape dissimilarity measure properties, such as metric and robustness properties, and of retrieval performance measures, and that of shape similarity measures.
- Similarity Fusion for Visual Tracking88
The approach contributes in two different aspects: multiple features are integrated into a unified similarity to enhance the discriminative ability of similarity measurements, and the neighborhood context of the samples in forthcoming frame are employed to further improve the measurements.
- 87
- 85
- 79
- Topologies for the digital spaces and78
It is shown that there are only two topologies in Z2 and five topologiesin Z3 whose connected sets are connected in the intuitive sense, and two of the fiveTopologies for Z3 are products of the topologies known from Z1 and Z2.
- 77
- Polygon Evolution by Vertex Deletion77
A simple approach to evolution of polygonal curves that is specially designed to fit discrete nature of curves in digital images, which leads to simplification of shape complexity with no blurring (i.e., shape rounding) effects and no dislocation of relevant features.
- Preserving Topology by a Digitization Process72
This paper derives conditions relating properties of real objects to the grid size of the sampling device which guarantee that a real object and its digital image are topologically equivalent, and implies that two digital images of a given object are topological equivalent.
- 71
- 71
- 70
- Regularized Diffusion Process for Visual Retrieval70
A new variant o diffusion process, which also operates on a tensor product graph is proposed, defined in three equivalent formulations (regularization framework, iterative framework and limit framework, respectively), which theoretically explain how this kind of diffusion process can better reveal the intrinsic relationship between objects.
- A Unified Curvature Definition for Regular, Polygonal, and Digital Planar Curves70
The proposed definition of visual curvature is the first ever that applies to regular curves as defined in differential geometry as well as to turn angles of polygonal curves and it yields stable curvature estimates of curves in digital images even under sever distortions.
- Optimal partial shape similarity70
The main contribution of this paper is the definition of an optimal partial shape similarity measure that works under these conditions and the presented novel approach to shape-based object recognition works even if only a small part of a given object is visible and the visible part is significantly distorted.
- DPNet: Dual-Path Network for Real-Time Object Detection With Lightweight Attention68
The experimental results demonstrate that DPNet achieves a state-of-the-art trade off between detection accuracy and implementation efficiency, and the dual-path architecture enables us to extract in parallel high-level semantic features and low-level object details.
- Robust Clustering as Ensembles of Affinity Relations68
The method can automatically select some points to form clusters, leaving other points un-grouped; thus it is inherently robust to large numbers of outliers, which has seriously limited the applicability of classical methods.
- Detection and recognition of contour parts based on shape similarity68
A system for detection and recognition of contour parts in digital images based on shape similarity to retrieve the most similar contours parts in a database of known contour segments.
- An elastic partial shape matching technique66
This work proposes to transform shapes into sequences and utilize an algorithm that determines a subsequence of a target sequence that best matches a query and demonstrates that the proposed algorithm outperforms the commonly used techniques in retrieval accuracy.
- No-reference mesh visual quality assessment via ensemble of convolutional neural networks and compact multi-linear pooling65
This work proposes a method based on deep learning for the mesh visual quality assessment without reference by extracting views from the 3D mesh and the corresponding mesh saliency, which is used to estimate the quality score.
- Discrete Representation of Spatial Objects in Computer Vision63
This presentation explains the development of the graph-based approach to embedding, and some of the techniques used in this approach have been described as “rocket science”.
- 61
- Multi-scale context for scene labeling via flexible segmentation graph59
A fusion model using flexible segmentation graph (FSG) is presented to explore multi-scale context for scene labeling problem, and yields similar to or better labeling result than competing models, and is also computationally efficient.
- Contour-based object detection as dominant set computation59
As the proposed method is able to get all the dominant sets, it can detect multiple objects in an image in one pass and determine an optimal scale of target object without a common enumeration of all possible scales.
- 56
- Discrete Skeleton Evolution53
The proposed pruning method allows us to overcome the instability of skeleton representation: a small boundary deformation leads to large changes in skeleton topology and is able to obtain very stable skeleton representation of planar shapes.
- 53
- Elastic Partial Matching of Time Series53
The proposed algorithm is an algorithm that automatically determines a subsequence b′ of a target time series b that best matches a query series a in a DAG (directed acyclic graph) to solve the problem of elastic matching of time series.
- SciER: An Entity and Relation Extraction Dataset for Datasets, Methods, and Tasks in Scientific Documents52
A new entity and relation extraction dataset for entities related to datasets, methods, and tasks in scientific articles, containing 106 manually annotated full-text scientific publications with over 24k entities and 12k relations is released.
- Skeleton pruning as trade-off between skeleton simplicity and reconstruction error52
This work proposes a simple algorithm to approximate the maximum of the Bayesian posterior probability which defines an order for iteratively removing the end branches to obtain the pruned skeleton and presents experimental results obtained without any parameter tuning.
- 52
- 52
- Particle filter with state permutations for solving image jigsaw puzzles51
This work proposes a novel inference approach in the sampling framework of Particle Filter to estimate the posterior density by exploring different orders of observations and selecting the most informative permutations of observations, which significantly broadens the scope of applications of the PF inference.
- 51
- Shape clustering: Common structure discovery49
This paper aims to address the problem of shape clustering by discovering the common structure which captures the intrinsic structural information of shapes belonging to the same cluster, based on a skeleton graph, named common structure skeleton graph (CSSG), which expresses possible correspondences between nodes of the individual skeletons of the cluster.
- 48
- 48
- Digitizations preserving topological and differential geometric properties47
Conditions which guarantee that every digitization process preserves important topological and differential geometric properties are presented, which allow us to determine the correct digitization resolution for a given class of real objects.
- Fusion with Diffusion for Robust Visual Tracking46
The advantages of the proposed approach to integrate multiple similarity measures is demonstrated on the task of visual tracking, where different aspects of the appearance similarity between the target object in frame t and target object candidates in frames t are integrated.
- 42
- 3D visual saliency and convolutional neural network for blind mesh quality assessment38
This work proposes a no-reference convolutional neural network (CNN) framework to estimate the perceived visual quality of 3D meshes and shows that the trained CNN achieves good rates in terms of correlation with human judgment and outperforms the most effective state-of-the-art methods.
- 38
- Learning Pixel-wise Alignment for Unsupervised Image Stitching37
An unsupervised image stitching framework, breaking through the coplanar constraints in homography estimation, yielding accurate pixel-wise alignment under limited overlapping regions, achieves state-of-the-art performance that surpasses both traditional and deep learning approaches by a large margin.
- SKELETON-BASED SHAPE CLASSIFICATION USING PATH SIMILARITY37
Using a new skeleton pruning method, it is able to obtain stable pruned skeletons even in the presence of significant contour distortions, and a new method for matching of skeleton graphs is proposed.
- Topological Repairing of 3D Digital Images35
A new randomized algorithm for repairing the topology of objects represented by 3D binary digital images, which finds application in repairing segmented images resulting from multi-object segmentations of other 3D digital multivalued images.
- Shape Matching for Robot Mapping35
A novel geometric model for robot mapping based on shape is presented, which allows for compact representations that are well-suited to bridge the gap from metric information needed in path planning to more abstract, i.e. topological or qualitative spatial knowledge desired in complex navigational tasks.
- Self-Supervised Learning with Masked Image Modeling for Teeth Numbering, Detection of Dental Restorations, and Instance Segmentation in Dental Panoramic Radiographs34
This study aims to utilize recent self-supervised learning methods like SimMIM and UM-MAE to increase the model efficiency and understanding of the limited number of dental radiographs.
- An open-source project for real-time image semantic segmentation33
This research presents a novel and scalable approach to integrate 3D image recognition and 3D audio and video recognition systems into the fabric of the internet.
- 33
- Shape Similarity and Visual Parts32
Human perception of shape is based on visual parts of objects to a point that a single, significant visual part is sufficient to recognize the whole object, which leads to two related problems of scale selection and subpart selection.
- Boosting Chamfer Matching by Learning Chamfer Distance Normalization31
A novel technique is proposed that significantly improves the performance of oriented chamfer matching on images with cluttered background and uses AdaBoost to learn a Normalized Oriented Chamfer Distance (NOCD).
- Topological Connectedness and 8-Connectedness in Digital Pictures31
This paper will prove that there is no topology on Z 2 in which the connected sets are exactly the 8-connected sets using different ideas, and the proof obtained will be much simpler and shorter.
- Human Detection Using Learned Part Alphabet and Pose Dictionary30
A compositional model for human detection in natural scenes using a discriminative alphabet, each grapheme of which is a mid-level element representing a body part, which correspond to various human poses and allow for robust, efficient matching.
- Contour based object detection using part bundles30
It is shown that appearance information can be used for improving detection for objects with distinctive texture when model contour does not sufficiently capture deformation of the objects.
- 30
- 30
- 29
- Rank-based self-training for graph convolutional networks28
This paper proposes a novel self-training approach through a rank-based model for improving the accuracy of GCNs on semi-supervised classification tasks, and proposes a rank aggregation of labeled sets obtained by different GCN models.
- 28
- Graph-based Approach28
In this approach, it is possible to show that some particular properties of a digital concept are the same as the corresponding properties of its continuous original.
- DMDD: A Large-Scale Dataset for Dataset Mentions Detection27
The dataset Mentions Detection Dataset (DMDD), the largest publicly available corpus for dataset mention detection, is introduced and the performance of various models on DMDD is analyzed to identify open problems in datasets mention detection.
- Graph Convolutional Networks based on manifold learning for semi-supervised image classification27
This work proposes a novel approach, the Manifold-GCN, based on GCNs for semi-supervised image classification, which is the first framework that allows the combination of GCNs with different types of manifold learning approaches for image classification.
- 27
- Shape Description and Search for Similar Objects in Image Databases26
A cognitively motivated similarity measure is presented and its properties are analyzed with respect to retrieval of similar objects in image databases of silhouettes of 2D objects and the experimental results justify that the shape matching procedure gives an intuitive shape correspondence and is stable withrespect to noise distortions.
- Restart and Random Walk in Local Search for Maximum Vertex Weight Cliques with Evaluations in Clustering Aggregation25
This paper proposes to use the restart and the random walk strategies to improve local search for MVWC and demonstrates that the solver outperforms the state-of-the-art for solving the derived MVWC problem and helps improve the final clustering results.
- Unsupervised object region proposals for RGB-D indoor scenes25
A novel unsupervised framework for automatically generating bottom up class independent object candidates for detection and recognition in cluttered indoor environments and is even able to outperform supervised state-of-the-art algorithms on the challenging NYU-v2 RGB-Depth dataset.
- 25
- 25
- ONLINE MULTIPLE TARGETS DETECTION AND TRACKING FROM MOBILE ROBOT IN CLUTTERED INDOOR ENVIRONMENTS WITH DEPTH CAMERA24
The proposed approach gives a novel framework for detection and tracking in indoor environments, and extends it to robot navigation by tracking, using a virtual top view created from an RGB-D camera named ground plane projection (GPP).
- 24
- Supportedness and tameness differentialless geometry of plane curves24
In order to relate boundary properties of discrete objects obtained by segmenting digital images to the corresponding properties of their continuous originals, the theory of tame arcs is based on concepts that can be directly transferred from the continuous to the discrete domain.
- Contour-based shape similarity24
A similarity measure for silhouettes of 2D objects is presented, and its properties are analyzed with respect to retrieval of similar objects in an image database and compared to well-known approaches in the literature.
- 23
- Integration of Single-view Graphs with Diffusion of Tensor Product Graphs for Multi-view Spectral Clustering23
A novel graph-based approach to integrate multiple representations to improve clustering performance and shows that the proposed approach outperforms state-of-the-art methods on benchmark datasets.
- Image retrieval and reversible illumination normalization23
This paper proposes a novel approach to retrieve similar images from image databases that works in the presence of significant illumination variations by using a reversible illumination normalization transformation.
- Building Polygonal Maps from Laser Range Data23
A new approach to the problem of building a global map from laser range data, utilizing shape based object recognition techniques originally developed for tasks in com- puter vision, which provides a link between the necessary low-level and the desired high-level information in robot navigation.
- The choice of vantage objects for image retrieval23
This paper suggests a loss measure to assess the quality of a set of vantage objects: for every image, a matching image is selected from the remaining images of the database by use of the vantage set, and the resulting dissimilarities are average the resulting Dissimilarities.
- Multidimensional Scaling on Multiple Input Distance Matrices22
This paper proposes a new algorithm called Multi-View Multidimensional Scaling (MVMDS) by considering each input distance matrix as one view, which can learn the weights of views automatically by exploring the consensus information and complementary nature of views.
- Special issue: Shape Representation and Similarity for Image Databases22
The emphasis in this Special Issue is on universal, non-parametric shape representations, and methods that are not projective invariant in the mathematical sense, but which achieve a reasonable robustness to projective transformation without any additional computation.
- Multicolor well-composed pictures22
This paper presents a solution which guarantees avoidance of the connectivity paradoxes related to the Jordan Curve Theorem for all multicolor pictures and implies that properties of algorithms used in computer vision can be stated and proved in a clear way, and that the algorithms themselves become simpler and faster.
- 21
- 20
- Clustering Aggregation as Maximum-Weight Independent Set20
This work formulate clustering aggregation as a special instance of Maximum-Weight Independent Set (MWIS) problem, and proposes a variant of simulated annealing method that takes advantage of this special structure.
- An Algorithm for a 3D Simplicity Test20
The first 3D algorithm for determining the simplicity of any three-color deletion configuration is presented, which is memory efficient and fast, and can be a very useful tool for designing 3D connectivity preserving thinning (or shrinking) algorithms.
- Semi-proximity continuous functions in digital images20
This work uses semi-proximity spaces to establish a formal relationship between the ''topological'' concepts of digital image processing and their continuous counterparts in @?^n and shows that semi- Proximity continuous functions are a useful tool in the difficult task of shape description.
- Orientation and Qualitative Angle for Spatial Reasoning.20
It is shown that the augmentation of arrangement knowledge by qualitative angles leads to interesting and useful inferences about relative positions of points in the plane.
- 19
- Making 3D binary digital images well-composed19
A new randomized algorithm for making 3D binary digital images that are not well-composed into well-Composed ones is introduced, the complexity and convergence of the algorithm are analyzed, and experimental evidence of its effectiveness when faced with practical medical imaging data is presented.
- Efficient Rank-Based Diffusion Process with Assured Convergence18
An efficient Rank-based Diffusion Process which combines both approaches and avoids the drawbacks of each one is proposed, capable of efficiently approximating a diffusion process by exploiting rank-based information, while assuring its convergence.
- Efficient Local Search for Minimum Dominating Sets in Large Graphs18
A new local search algorithm ScBppw (Score Checking and Best-picking with Probabilistic Walk) to solve the MinDS problem in large graphs, which significantly outperforms state-of-the-art MinDS solvers.
- Deep Context Convolutional Neural Networks for Semantic Segmentation18
This paper designs a multi-scale deep context convolutional network (MDCCNet), which combines the feature maps from different levels of network in a holistic manner for semantic segmentation and demonstrates that the method outperforms or is comparable to state-of-the-art methods on PASCAL VOC 2012 and SIFTFlow semantic segmentations datasets.
- 18
- Relevance Ranking of Video Data Using Hidden Markov Model Distances and Polygon Simplification18
Using these techniques, a viewer can browse a video at the level of summarization that suits his patience level, and applications include the creation of a smart fast-forward function for digital VCRs, and the automatic creation of short summaries that can be used as previews before videos are downloaded from the web.
- 17
- Size Adaptive Selection of Most Informative Features17
A novel method to select the most informatives subset of features, which has little redundancy and very strong discriminating power is proposed, and the obtained classification results significantly outperform the state-of-the-art results on most test datasets.
- Geometric Robot Mapping17
A technique to create a global map of a robot's surrounding by converting the raw data acquired from a scanning sensor to a compact map composed of just a few generalized polylines (polygonal curves).
- Transductive Domain Adaptation with Affinity Learning16
It is demonstrated that with improved similarities between source and target data, spectral embedding provides a better data representation, which boosts the prediction accuracy, and the effectiveness of the proposed method is validated on standard benchmark datasets for visual object recognition.
- New EM derived from Kullback-Leibler divergence16
A new EM framework is introduced in which it is possible not only to optimize the model parameters but also the number of model components, and a key feature of this approach is that it uses nonparametricdensity estimation to improve parametric density estimation in the EM framework.
- Coupling Deep Textural and Shape Features for Sketch Recognition15
This paper explicitly explores the shape properties of sketches, which has almost been neglected before in the context of deep learning, and proposes a sequential dual learning strategy that combines both shape and texture features.
- A Realistic Digitization Model of Straight Lines15
A realistic mathematical model of a digitized edge which handles both blurring and arbitrary thresholding is presented and it is shown that a thresholded digital image of a blurred half-plane obtained for some unknown threshold value is equal to theimage of a perfectly focused half-planes with the same slope obtained by object boundary quantization.
- Scene Parsing Via Dense Recurrent Neural Networks With Attentional Selection14
This paper proposes dense RNNs for scene labeling by exploring various long-range semantic dependencies among image units and develops an end-to-end scene labeling system integrating with convolutional neural networks (CNNs).
- Learning adaptive contrast combinations for visual saliency detection14
This paper introduces a novel saliency detection method based on weighted linear multiple kernel learning (WLMKL) framework, which is able to adaptively combine different contrast measurements in a supervised manner, yielding more robust and accurate saliency maps.
- Weakly Supervised Shape Based Object Detection with Particle Filter14
An efficient approach to construct shape models composed of contour parts with partially-supervised learning that can easily transfer parts structure to different object classes as long as they have similar shape is described.
- Contour Grouping Based on Contour-Skeleton Duality14
A method for grouping relevant object contours in edge maps by taking advantage of contour-skeleton duality and adapting the state-of-the-art probabilistic framework namely Rao-Blackwellized particle filtering that has been successfully applied to SLAM is presented.
- Symmetry of Shapes Via Self-similarity14
A simple and novel approach to identify main similarity axes by maximizing self-similarity of object contour parts divided by the axes by finding two subsets of the critical points such that the two contours determined by the subsets are maximally similar.
- 14
- Digital geometric methods in document image analysis14
This paper derives conditions relating properties of real objects to the grid size of the sampling device which guarantee that a real object and its digital image are topologically equivalent and proves that a topologically invariant digitization must result in well-composed or strongly connected sets and that only certain local neighborhoods are realizable for such a digitization.
- Continuity of the discrete curve evolution14
It is proved that the process of the discrete curve evolution is continuous: if polygon Q is close to polygon P, then the polygons obtained by their evolution remain close.
- 13
- Learning Graph Convolutional Network for Blind Mesh Visual Quality Assessment13
A new method for blind mesh visual quality assessment (MVQA) based on a graph convolutional network that addresses the node classification problem to predict the perceived visual quality and investigates several perceptual and visual features to select the most effective combination.
- 13
- Densifying Distance Spaces for Shape and Image Retrieval13
This work presents an innovative approach that adds synthetic points directly to distance spaces, and calls these synthetic points ghost points since they are not represented by vectors of features, and consequently, cannot be directly visualized.
- Multi robot mapping using force field simulation13
The presented algorithm solves the alignment problem of such laser scans utilizing a gradient descent approach motivated by physics, namely simulation of movement of masses in gravitational fields, but exchanges laws of physics with constraints given by human perception.
- 13
- Digitizations preserving shape13
It is shown in this paper how it is possible to digitize a wide class of planar compacta in such a way that their digitizations have the same shape, in the sense of Borsuk, as the original set.
- 12
- Rank Flow Embedding for Unsupervised and Semi-Supervised Manifold Learning12
This work proposes a novel manifold learning algorithm named Rank Flow Embedding (RFE), based on ideas recently exploited by manifold learning approaches, which include hypergraphs, Cartesian products, and connected components, which computes context-sensitive embeddings, which are refined following a rank-based processing flow.
- DPNET: Dual-Path Network for Efficient Object Detection with Lightweight Self-Attention12
This paper presents a dual path network, named DPNet, for efficient object detection with lightweight self-attention, which achieves promising detection results on the COCO dataset.
- Training convolutional neural network from multi-domain contour images for 3D shape retrieval12
This work takes 2D sketches and edge maps of 2D views of 3D models as a unified domain to train the Convolutional Neural Network (CNN), and shows that the CNN can successfully learn the object structure from different types of clues.
- Multiscale Random Fields with Application to Contour Grouping12
A new interpretation of multiscale random fields (MSRFs) that admits efficient optimization in the framework of regular (single level)random fields (RFs), based on a new operator, called append, that combines sets of random variables (RVs) to single RVs.
- Using spatiotemporal blocks to reduce the uncertainty in detecting and tracking moving objects in video12
The motion detection and tracking experiments demonstrate that the novel method although simpler than a state-of-the-art method based on the Stauffer-Grimson Gaussian mixture model has superior performance.
- 12
- Shape-Based Robot Mapping12
The described shape based approach is an improvement of the underlying geometric models of todays SLAM implementations and allows for compact representations that are well-suited to bridge the gap from metric information needed in robot motion and path planning to more abstract, i.e. topological or qualitative spatial knowledge desired in complex navigational tasks or communication.
- 12
- 11
- 11
- 11
- Face recognition via fast dense correspondence11
An improved sparse-representation based face recognition method, which constructs dense pixel correspondences between training and testing facial samples and a matching energy function is designed to perform face alignment based on dense pixel correspondence to address the issue of non-rigid deformations.
- Face identification using reference-based features with message passing model11
A system for face identification that adopts the one-shot similarity kernel for learning the similarity of two face images and proposes a graph-based method for selecting an optimal set of reference images.
- Analysis of Facial Images across Age Progression by Humans11
A thorough study of human ability on two tasks, face verification and age estimation, for facial images taken at different ages, which helps understanding roles of different factors including age group, age gap, race, and gender.
- Skeleton Pruning by Contour Partitioning11
This paper establishes a unique correspondence between skeleton branches and subarcs of object contours and shows that this approach preserves skeleton topology, does not shift the skeleton, and it does not shrink the remaining branches.
- Discrete Approach to Curve Evolution11
It is shown that the proposed approach to evolution of digital planar curves by digital linearization leads to simplification of shape complexity, analog to evolutions guided by diffusion equations, but with no blurring (i.e., shape rounding) effects and no dislocation of relevant features.
- Generalized convexity: CP3 and boundaries of convex sets11
It is proved that a closed curve is the boundary of a convex set, and a simple arc is part of the boundary, iff they have property CP 3 .
- Direct Visual Grounding by Directing Attention of Visual Tokens10
A novel loss function is proposed that directly supervises the attention of visual tokens to corresponding language tokens in images and directly grounds the answer language tokens in images by directing their attention to the relevant visual tokens.
- 10
- Sequential Monte Carlo for Maximum Weight Subgraphs with Application to Solving Image Jigsaw Puzzles10
The inference quadruples the accuracy of the puzzle assembly compared to that of loopy belief propagation and proposes a novel SMC algorithm for obtaining maximum a posteriori estimate of a high-dimensional posterior distribution.
- Maximal Cliques that Satisfy Hard Constraints with Application to Deformable Object Model Learning10
A novel inference framework for finding maximal cliques in a weighted graph that satisfy hard constraints is proposed and yields discriminative part based object models that achieve very good detection rate, and outperform other methods on object classes with large deformation.
- Robot mapping for rescue robots10
The detection of shape similarity is the key to generate global maps even of highly cluttered environments and thus the proposed approach is particularly suitable for applications in the field of rescue robots.
- 10
- Shape Similarity Based on the Qualitative Spatial Reasoning Calculus eOPRAm9
This paper investigates the use of qualitative spatial representations QSR about relative direction and distance for shape representation using the conceptual neighborhood to directly establish a conceptual neighborhood between shapes that translates into a similarity metric for shapes.
- Locality Preserving Projection for Domain Adaptation with Multi-Objective Learning9
A novel locality preserving projection method for domain adaptation task, which can find a linear mapping preserving the 'intrinsic structure' for both source and target domains, and shows better prediction accuracy.
- 9
- Activity and Motion Detection Based on Measuring Texture Change9
This work estimates the speed of texture change by measuring the spread of texture vectors in their feature space and proposes techniques to detect increased activities in videos by learning a normal amount of textures change over time.
- 9
- Activity and Motion Detection Based on Measuring Texture Change9
This work estimates the speed of texture change by measuring the spread of texture vectors in their feature space and proposes techniques to detect increased activities in videos by learning a normal amount of textures change over time.
- 9
- A rotation robust shape transformer for cartoon character recognition8
A rotation robust shape transformer for cartoon character recognition, which beats both handcrafted and deep learning methods on the proposed challenging Cartoon dataset and traditional datasets, and develops an online cartoon character recognition application for animation scenarios.
- Affinity Learning for Mixed Data Clustering8
A novel affinity learning based framework for mixed data clustering, which includes: how to process data with mixed-type attributes, how to learn affinities between data points, and how to exploit the learned affinITIES for clustering.
- 8
- 8
- 7
- Enhanced Masked Image Modeling for Analysis of Dental Panoramic Radiographs7
A novel self-distillation enhanced self-supervised learning on top of the masked image modeling (SimM IM) Transformer, called SD-SimMIM, to improve the outcome with a limited number of dental radiographs, and shows that it outperforms other self- supervised learning methods.
- Spherical coverage verification7
Using the reduction of non-degenerated concave quadratic programming (QP) problem, it is demonstrated that spherical coverage verification is NP hard and a recursive algorithm based on reducing the problem to several lower dimension subproblems is proposed.
- Small moving targets detection using outlier detection algorithms7
This paper compares the performance of two incremental outlier detection algorithms, namely the incremental connectivity-based outlier factor and the incremental local outlier factors to modified Stauffer-Grimson algorithm.
- SKELETONIZATION WITH PARTICLE FILTERS7
Experimental results demonstrate that the proposed method significantly outperforms other well-known methods for skeleton computation and an extension of the method to computing skeletons of gray level images is proposed and provides promising experimental results.
- Shape Classification Based on Skeleton Path Similarity7
Using a skeleton pruning method, this work is able to obtain stable pruned skeletons even in the presence of significant contour distortions, and presents excellent classification results for complete shape.
- 7
- 7
- 7
- Towards a Generalization of Self-localization6
This chapter proposes an approach to identify locations on a high, abstract level of representation to recognize locations described as configurations of shape features and evaluates this approach in comparison to other approaches in a self-localization task and a generalized localization task based on a schematic map.
- 6
- Shape Decomposition and Shape Similarity Measure6
A simple and natural rule for decomposition of 2D objects into parts of visual form based on a novel curve evolution method by digital linearization in which a significant visual part will become a convex part at some level of the evolution.
- An indexing technique for implementing command relations6
This paper presents an indexing technique which allows us to implement a simple and efficient check for most cases of command relations, and shows a wide perspective of applications in the implementation of other linguistic phenomena in syntax as well as in semantics.
- 5
- 5
- Common Object Discovery as Local Search for Maximum Weight Cliques in a Global Object Similarity Graph5
The proposed approach is evaluated on the PASCAL VOC image dataset and the YouTube-Objects video dataset, and it demonstrates superior performance over recent state-of-the-art approaches.
- Stable Feature Selection with Minimal Independent Dominating Sets5
The proposed MIDS framework complements standard feature selection algorithms like SVM-RFE, stability lasso and ensemble SVM RFE and achieves better prediction accuracy, as the experimental results clearly demonstrate.
- Performance of 6 D LuM and FFS SLAM5
The question addressed in this paper is, if different ways of evaluating the performance of SLAM algorithms project different strengths and how can the evaluations be useful in selecting an algorithm.
- 5
- 5
- 5
- 5
- Recovering a Polygon from Noisy Data5
A vertex elimination process is defined that accomplishes the recovery of an unknown polygon from noisy digital data obtained by digitizing either an image of the (solid) polygon or a sequence of points on its boundary by selecting a subset of the data points as vertices.
- Automatic Extraction of Relevant Frames from Videos by Polygon Simplification5
This work presents a polygon simplification method that works in multidimensional spaces or even in semi-metric spaces that need not be vector spaces, and requires only that a (semi-)distance between pairs of points be defined that needs not satisfy the triangle inequality.
- Self-Supervised Masked Deep Embeddings for Dental Caries Detection4
This research paves the way for AI-driven solutions in dentistry, offering valuable tools for dentists and patients in the early detection and management of dental caries, and suggests that self-supervised learning can be effectively applied to other areas of medical imaging where data scarcity is a significant challenge.
- 4
- 4
- 4
- 4
- 4
- Learning Non-linear Calibration for Score Fusion with Applications to Image and Video Classification4
- 4
- Contour Extraction Using Particle Filters4
A novel approach to extract object region from an image by tracking the enclosing contour by utilizing a particle filter framework and utilizing the filter to track the optimal contour path pixel by pixel.
- 4
- 4
- 4
- 4
- 4
- Invariant Thinning and Distance Transform4
The aim of this paper is to show that under certain assumptions a ‘classical’ thinning method, which has invariance with respect to motions as a distinctive feature, also can be interpreted as a distance-transform-based method.
- Conditions that Guarantee a Digitization Process Preserves Topology4
Conditions relating properties of real objects to the grid size of the sampling device guarantee that a real object and its digital image are topologically equivalent.
- VAPO: Visibility-Aware Keypoint Localization for Efficient 6DoF Object Pose Estimation3
This paper construct VAPO (Visibility-Aware POse estimator) by integrating the visibility-aware importance with a state-of-the-art pose estimation algorithm, along with additional positional encoding, and derives real-valued visibility-aware importance from binary labels based on the PageRank algorithm.
- Rank-based Hashing for Effective and Efficient Nearest Neighbor Search for Image Retrieval3
This work proposes an unsupervised data-independent approach for nearest neighbor searches, which can be used with different features, including deep features trained by transfer learning, and uses a rank-based formulation and exploits a hashing approach for efficient ranked list computation at query time.
- Image Retrieval with Similar Object Detection and Local Similarity to Detected Objects3
This work localizes the most similar region in the database image and only this region is used for computing image similarity and achieves significantly better retrieval performance than the off-the-shelf object localization-based retrieval methods and end-to-end trained triplet method with a region proposal network.
- 3
- Location-Aware Image Classification3
This paper proposes a novel location-aware image classification framework that outperforms the popular spatial pyramid matching SPM method and the Region Based Latent SVM RBLSVM method on the challenging PASCAL VOC dataset.
- Unsupervised Segmentation of RGB-D Images3
An unsupervised segmentation algorithm that is carefully crafted to balance the contribution of color and depth features in RGB-D images and yields comparable results to recently proposed supervised segmentation methods on challenging NYU Depth dataset v2.
- 3
- Force Field Based n-Scan Alignment.3
The presented algorithm solves the alignment problem utilizing a gradient descent approach motivated by physics, but exchanges laws of physics with constraints given by human perception.
- 3
- Image Similarity Measures for Video Processing3
It is shown that a simple extension of image histograms to include the position information of the centroids of histograms bins leads to a useful representation for video analysis.
- 3
- 3
- 3
- 3
- 3
- Well-composedness of digital sets3
A special class of subsets of binary digital images called `well-composed sets' are defined; for example, the Jordan Curve Theorem holds, the Euler characteristic is locally computable, and there is only one connectedness relation, since 4- and 8-connectedness are equivalent.
- 2
- 2
- Image Retrieval with Self-Supervised Divergence Minimization and Cross-Attention Classification2
This work uses self-supervision with a novel divergence loss framework alongside a simple data flow adjustment that minimizes a distribution over a database directly during training to show that jointly learning a query representation over a database is a competitive and often improved alternative to traditional contrastive methods for image retrieval.
- 2
- 2
- Weighted Linear Multiple Kernel Learning for Saliency Detection2
This paper presents a novel saliency detection method based on weighted linear multiple kernel learning (WLMKL), which is able to adaptively combine different contrast measurements in a supervised manner and produces more robust and accurate saliency maps.
- 2
- Polygonal Approximation of Point Sets2
The presented approach uses local support estimation motivated by human visual perception to evaluate support in data points of EM components after each EM step and is able to recognize a locally optimal solution that is not globally optimal, and modify the number of model components and their parameters.
- 2
- Navigation toward Non-static Target Object Using Footprint Detection Based Tracking2
A Footprint Detection based Tracker (FD-Tracker) is proposed to robustly track the target object in such scenarios and the results demonstrate that using footprint can boost the performance of visual tracker.
- View-Invariant Object Detection by Matching 3D Contours2
The proposed approach for view-invariant object detection directly in 3D with following properties significantly outperforms the state-of-the-art 2D approaches, in particular, latent SVM object detector, as well as recently proposed approaches for object detection in RGB-D data.
- Diffusion on a Tensor Product Graph for Semi-Supervised Learning and Interactive Image Segmentation2
It is shown that SADP converges to a closed-form solution by proving its equivalence to a diffusion process on a tensor product graph, and the proposed approach can be used in interactive image segmentation, which is also called semi-supervised image segmentsation.
- Distance Learning Based on Convex Clustering2
In the proposed framework, the data is transformed from elongated structures into compact ones by a novel distance learning algorithm and a convex clustering algorithm is used to cluster the transformed data.
- Computing Stable Skeletons with Particle Filters2
Experimental results are provided that clearly demonstrate that the proposed method significantly outperforms other well-known methods for skeleton computation.
- Polygonal Approximation of Point Sets2
The presented approach uses local support estimation motivated by human visual perception to evaluate support in data points of EM components after each EM step and is able to recognize a locally optimal solution that is not globally optimal, and modify the number of model components and their parameters.
- Motion detection and object tracking in grayscale videos based on spatiotemporal texture changes2
The proposed spatiotemporal motion detection method does not require any post-processing, which is an essential step required by the Gaussian mixture model, and real-time software designed to detect motion based on the spatiotemporal texture change, is evaluated using color and infrared cameras.
- Reliability of motion features in surveillance videos2
Results on benchmark real-life videos demonstrate the capability of the proposed techniques to successfully eliminate problems due to change in light conditions, transition/compression artifacts and unwanted camera motions.
- Instantaneous reliability assessment of motion features in surveillance videos2
The main strategy is to estimate the feature properties when the features are reliable computed, so that any set of features that does not have these properties is detected as being unreliable.
- 2
- 2
- 2
- 2
- Toward non-parametric digital shape representation and recovery2
It is shown that if an object is parallel regular, very few digital boundary patterns are realizable and that each such digital neighborhood has a well-defined geometric interpretation with respect to tangent direction, and described how a generalized torus can be recovered from a single intensity image using digital non-parametric models and methods.
- Connection relations and quantifier scope2
A formalism will be presented in this paper which makes it possible to realise the idea of assigning only one scope-ambiguous representation to a sentence that is ambiguous with regard to quantifier scope.
- Scaling Performance and Low-Resource Annotation with Many-Shot In-Context Learning for Named Entity Recognition1
Scaling to hundreds of in-context examples enables LLMs to match or even surpass the performance of fully supervised BERT models, and many-shot in-context annotation can generate high-quality labeled data, leading to approximately 10% absolute F1 improvement over existing state-of-the-art approaches when used to fine-tune BERT on low-resource NER.
- Self-Supervised Learning of Deep Embeddings for Classification and Identification of Dental Implants1
The proposed Masked Deep Embedding (MDE) pre-training method, extending the masked autoencoder (MAE) transformer, significantly enhances dental implant detection performance compared to baselines, achieving substantial performance improvement in implant detection.
- Querying Climate Knowledge: Semantic Retrieval for Scientific Discovery1
This paper introduces a domain-specific Knowledge Graph built from climate publications and broader scientific texts, aimed at improving how climate knowledge is accessed and used, and demonstrates how the KG answers questions using Cypher queries.
- 1
- 1
- 1
- DRBANET: A Lightweight Dual-Resolution Network for Semantic Segmentation with Boundary Auxiliary1
This paper introduces a lightweight dual-resolution network, called DRBANet, aiming to refine semantic segmentation results with the aid of boundary information, and achieves promising trade-off between segmentation accuracy and running efficiency.
- BASNet: improving semantic segmentation via boundary-assistant symmetrical network1
A novel symmetrical network, called BASNet, which contains four components: the pre-trained ResNet-101 backbone, semantic segmentation branch (SSB), boundary detection branch (BDB), and aggregation module (AM).
- 1
- 1
- Semi-Supervised Learning on an Augmented Graph with Class Labels1
This paper proposes a diffusion-based affinity learning algorithm on an augmented graph consisting of not only the nodes of labeled and unlabeled data but also artificial nodes representing class labels that achieves superior results on many standard data sets.
- 1
- 1
- Optimal subsequence bijection and classification of imbalanced data sets1
This work proposes using an algorithm that determines the optimal subsequence bijection (OSB) of a query and target time series, and makes several significant improvements to the original OSB algorithm.
- 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
- 1
- Evaluating Reliability of Motion Features in Surveillance Videos1
The claim is that the fact that the computed features will never be 100% reliable, and focus the attention on computing reliability measures, means for an intelligent system for video analysis that in addition to feature computation, it should perform instantaneous evaluation of their reliability, and adapt its behavior in accordance to the reliability.
- A Two-Stream Approach for Priority Management and Adaptive Rate Control in Multimedia Applications1
A two-stream approach for adaptive rate control in multimedia applications that keeps track of the available bandwidth of the network path and dynamically adjust the sending rate of the traffic stream close to the optimal rate.
- 1
- Digital geometric methods in image analysis and compression1
The focus of this paper is developing digital geometric models and methods for finding regular structure in digital document images, and the applicability of the digital geometric approach is also demonstrated on images taken under orthographic and perspective projection.
- Topologically invariant methods in document image analysis1
Using the derived topological model of a well-composed digital image, the effectiveness of this model with respect to the digitization, thresholding, correction, and compression of digital document images is demonstrated.
- 1
- 1
- –
- –
- CypherLens: An Interactive Demo for Evaluating and Diagnosing NL-to-Cypher Systems–
CypherLens enables users to move beyond binary evaluation toward a more comprehensive and interpretable analysis of NL-to-Cypher systems, supporting debugging, failure analysis, and model comparison.
- Enhancing Renal Tumor Malignancy Prediction: Deep Learning with Automatic 3D CT Organ Focused Attention–
A deep learning framework was developed utilizing an Organ Focused Attention (OFA) loss function to modify the attention of image patches so that organ patches attend only to other organ patches, suggesting that this approach offers a more efficient and reliable method for malignancy prediction, thereby enhancing clinical decision-making in renal cancer diagnosis.
- ERUnderstand: Evaluating Vision-Language Models on Structured ER Diagrams–
Evaluating state-of-the-art Vision-Language Models (VLMs), it is found that while common ERD elements are recovered reliably, performance drops sharply on weak entities, and reasoning-augmented models improve overall performance by 15-25% but remain sensitive to linguistic priors and increasing diagram complexity.
- –
- –
- Learning Object Focused Attention–
An adaptation to the training of Vision Transformers that allows for an explicit modeling of objects during the attention computation is proposed by adding a new branch to selected attention layers that computes an auxiliary loss which is called the object-focused attention (OFA) loss.
- Iterated Clique Reductions in Vertex Weighted Coloring for Large Sparse Graphs–
A theorem is indicated which indicates that the reduction effects of the proposed reduction algorithm are equivalent to that of a counterpart which enumerates all maximal cliques in the whole graph if the run time is sufficiently long.
- Weakly Supervised Annotations for Multi-modal Greeting Cards Dataset–
This work proposes to aggregate features from pretrained images and text embeddings to learn abstract visual concepts from GCD, and demonstrates that the proposed dataset is also useful for generating greeting card images using pre-trained text-to-image generation model.
- –
- –
- –
- –
- –
- –
- –
- Matching of 3D Objects Based on 3D Curves–
A novel approach to 3D object retrieval capable to match query objects generated by a user with those captured by a depth device (RGB-D) and all the information collected to describe the object is used for matching.
- Corrigendum to “A tutorial explaining a machine vision model that emulates human performance when it recovers natural 3D scenes from 2D images” [J. Math. Psychol. 56 (2012) 217–231]–
Corrigendum Corrigendum to ‘‘A tutorial explaining a machine vision model that emulates human performance when it recovers natural 3D scenes from 2D images’’.
- Graph-based inference with constraints for object detection and segmentation–
A novel framework for contour based object detection, by replacing the hough-voting framework with finding dense subgraph inference, and a novel shape matching scheme suitable for partial matching of edge fragments is proposed.
- –
- Binarization of Gray-Level Images Based on Skeleton Region Growing–
A new binarization method of gray level images that performs well on images with low contrast, noise and non-uniform illumination and is tested on different types of document components and degradations.
- Robotic Navigation Emulating Human Performance–
The mere fact that the robot can actually see a 3D scene veridically and plan its actions effectively within it provides evidence for the belief that its tools are at least biologically-plausible even if they ultimately prove to be different from those actually used in the human visual system.
- Shape similarity analysis of regions of interest in medical images–
A new representation technique of 2D contour shapes and a sequence similarity measure to characterize 2D regions of interest in medical images is introduced and it can be easily applied on medical diagnosis in all cases of which shape difference is an important clue.
- Hierarchical image segmentation for learning object priors–
The proposed segmentation approach naturally combines experience based and image based information and takes into account the region context information by averaging the probability maps over different levels of the hierarchical segmentation algorithm.
- –
- –
- Partial shape similarity of contours is needed for object recognition–
Psychophysical evidence is provided that recognition of parts of object contours is a necessary component of object recognition and existing shape similarity measures do not satisfy these requirements, and a new partial shape similarity measure is proposed.
- Mobile Robot Mapping and Immersive Building Simulation–
A framework for integrated Augmented Reality (AR) architecture for indoor thermal performance data visualization that utilizes a mobile robot to generate environment maps that will facilitate rapid multi-room mobile AR visualizations is discussed.
- Data Visualization by Pairwise Distortion Minimization–
Noise masks pairwise dissimilarities between data points and analogous latent quantities; this difference in approach allows us to build in some extra flexibility into the interpretation and modeling of high-dimensional data.
- A framework to create resource-bounded network services–
A two step technique to limit, with fixed arbitrary probability, demand for resources exceeding total available resources at a node, which can be applied to both component-based and ad-hoc service environments.
- –
- –
- –
- –
- –
- –
- Continuous Representations of Real Objects–
This class is too general to allow us to formally describe any shape properties of sets in this class and there are curves with very unnatural properties (e.g., plane filling curves).
- –
- –
- –
- –
- Digital topology of multicolor images–
A solution is presented which guarantees the authors avoid the connectivity paradoxes related to the Jordan Curve Theorem for all multicolor images, which implies that properties of algorithms used in computer vision can be stated and proved in a clear way, and that the algorithms themselves become simpler and faster.
- Topologie in Depiktionen–
Dabei kommen solche Begriffe wie Umgebung, Weg, geschlossene Kurve, Stetigkeit and Zusammenhang ins Spiel, die gutfundierte Definitionen in der Topologie, die unseren Intuitionen entsprechen.
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.