Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks244 from public data
- Studying aesthetics in photographic images using a computational approach1,056
This paper treats the challenge of automatically inferring aesthetic quality of pictures using their visual content as a machine learning problem, with a peer-rated online photo sharing Website as data source and extracts certain visual features based on the intuition that they can discriminate between aesthetically pleasing and displeasing images.
- 1Salient Object Detection: A Survey848
A comprehensive review of recent progress in salient object detection is provided and this field is situate among other closely related areas such as generic scene segmentation, object proposal generation, and saliency for fixation prediction.
- 454
- IRM441
The IRM measure for evaluating overall similarity between images incorporates properties of all the regions in the images by a region-matching scheme, which achieves more accurate retrieval at higher speed than several existing systems.
- 309
- Learning Open Set Network with Discriminative Reciprocal Points294
A new concept, Reciprocal Point, which is the potential representation of the extra-class space corresponding to each known category, which can indirectly introduce the unknown information into the learner with only known classes, so as to learn more compact and discriminative representations.
- Low-Resolution Face Recognition in the Wild via Selective Knowledge Distillation210
This paper proposes a learning approach to recognize low-resolution faces via selective knowledge distillation by selectively distil the most informative facial features from the teacher stream by solving a sparse graph optimization problem, which is then used to regularize the fine-tuning process of the student stream.
- Selectivity or Invariance: Boundary-Aware Salient Object Detection206
A novel boundary-aware network with successive dilation for image-based SOD, where the feature selectivity at boundaries is enhanced by incorporating a boundary localization stream, while the feature invariance at interiors is guaranteed with a complex interior perception stream.
- Transductive Episodic-Wise Adaptive Metric for Few-Shot Learning197
A Transductive Episodic-wise Adaptive Metric (TEAM) framework for few-shot learning is proposed, by integrating the meta-learning paradigm with both deep metric learning and transductive inference and leverages an attention-based bi-directional similarity strategy for extracting the more robust relationship between queries and prototypes.
- Pyramidal Feature Shrinking for Salient Object Detection189
The proposed pyramidal feature shrinking network (PFSNet), which aims to aggregate adjacent feature nodes in pairs with layer-by-layer shrinkage, so that the aggregated features fuse effective details and semantics together and discard interference information.
- 173
- IMAGINATION172
The proposed IMAGINATION (IMAge Generation for INternet AuthenticaTION), a system for the generation of attack-resistant, user-friendly, image-based CAPTCHAs, produces controlled distortions on randomly chosen images and presents them to the user for annotation from a given list of words.
- 167
- A Benchmark Dataset and Saliency-Guided Stacked Autoencoders for Video-Based Salient Object Detection154
The proposed unsupervised baseline approach for video-based SOD is compared with 31 state-of-the-art models on the proposed dataset and outperforms 30 of them, including 19 image-based classic (unsupervised or non-deep learning) models, six image- based deep learning models, and five video- based unsuper supervised models.
- Pyramid Grafting Network for One-Stage High Resolution Saliency Detection147
An attention-based Cross-Model Grafting Module (CMGM) is proposed to en-able CNN branch to combine broken detailed information more holistically, guided by different source feature during decoding process, and an Attention Guided Loss (AGL) to explicitly supervise the attention matrix generated by CMGM to help the network better interact with the attention from different models.
- 141
- Fast Discrete Distribution Clustering Using Wasserstein Barycenter With Sparse Support138
A modified Bregman ADMM approach for computing the approximate discrete Wasserstein barycenter of large clusters that achieves high accuracy empirically at a much reduced computational cost and is highly competitive in comparison with some widely used methods in the corresponding areas.
- Complementary Trilateral Decoder for Fast and Accurate Salient Object Detection137
A novel framework to treat semantic context, spatial detail and boundary information separately in the decoder part of the Complementary Trilateral Decoder for saliency detection with three branches: Semantic Path, Spatial Path and Boundary Path is proposed.
- Collaboration over time: characterizing and modeling network evolution124
A novel stochastic model, Stochastic Poisson model with Optimization Tree (Spot) is proposed to efficiently predict any increment of collaboration based on the local neighborhood structure and empirical results show that Spot outperforms Support Vector Regression by better fitting collaboration records and predicting the rate of collaboration.
- Transformer-based Dual Relation Graph for Multi-label Image Recognition120
This paper proposes a novel Transformer-based Dual Relation learning framework, constructing complementary relationships by exploring two aspects of correlation, i.e., structural relation graph and semantic relation graph, which aims to capture long-range correlations from object context by developing a cross-scale transformer-based architecture.
- 116
- Health-Related Hot Topic Detection in Online Communities Using Text Clustering111
Automatic topic detection based on document clustering is an alternative approach for extracting health-related hot topics in online communities and is integrated into medical domain-specific features to represent the messages posted in online health communities.
- 105
- 95
- 91
- 83
- Is Depth Really Necessary for Salient Object Detection?82
This work proposes a novel depth-aware salient object detection framework with only RGB information as input for inference, which not only surpasses the state-of-the-art performance on five public RGB SOD benchmarks, but also exceeds the RGBD-based methods on five benchmarks by a large margin, while adopting less information and implementation light-weighted.
- Clustering Based on a Multilayer Mixture Model82
This article explores a clustering approach that models each cluster by a mixture of normals, and develops algorithms to estimate the model and perform clustering based on the classificationmaximum likelihood (CML) and mixture maximum likelihood (MML) criteria.
- 81
- Part-Guided Relational Transformers for Fine-Grained Visual Recognition80
This framework, namely PArt-guided Relational Transformers (PART), is proposed to learn the discriminative part features with an automatic part discovery module, and to explore the intrinsic correlations with a feature transformation module by adapting the Transformer models from the field of natural language processing.
- Cartoon Face Recognition78
This work presents a new challenging benchmark dataset, consisting of 389,678 images of 5,013 cartoon characters annotated with identity, bounding box, pose, and other auxiliary attributes, and proposes a multi-task domain adaptation approach that jointly utilizes the human and cartoon domain knowledge with three discriminative regularizations.
- Efficient Low-Resolution Face Recognition via Bridge Distillation78
By learning low- resolution face representations and mimicking the adapted high-resolution knowledge, a light-weight student model can be constructed with high efficiency and promising accuracy in recognizing low-resolution faces.
- Look Through Masks: Towards Masked Face Recognition with De-Occlusion Distillation78
This work proposes to migrate the mechanism of amodal completion for the task of masked face recognition with an end-to-end de-occlusion distillation framework, which consists of two modules that applies a generative adversarial network to perform face completion.
- Salient Object Detection With Purificatory Mechanism and Structural Similarity Loss77
A novel convolutional neural network with purificatory mechanism and structural similarity loss to better locate preliminary salient objects and purify results from hard-to-distinguish regions, thus accurately predicting the locations and details of salient objects.
- Robust multiple cameras pedestrian detection with multi-view Bayesian network71
A robust multiple cameras pedestrian detection approach with multi-view Bayesian network model (MvBN) that outperforms several state-of-the-art algorithms remarkably and demonstrates high robustness in different surveillance scenes is proposed.
- 69
- DehazeFlow: Multi-scale Conditional Flow Network for Single Image Dehazing68
DhazeFlow is proposed, a novel single image dehazing framework based on conditional normalizing flow that surpasses the state-of-the-art methods in terms of PSNR, SSIM, LPIPS, and subjective visual effects.
- Frequency Representation Integration for Camouflaged Object Detection66
This paper rethink COD from the perspective of frequency components and proposes a Frequency Representation Integration Network to mine informative cues from them and surpasses existing state-of-the-art methods by a large margin.
- 66
- 65
- Dual Adaptive Representation Alignment for Cross-Domain Few-Shot Learning59
This work proposes an effective dual adaptive representation alignment approach to address the cross-domain few-shot learning problem where only extremely few samples are available in target domains and focuses on the fast adaptation capability of meta-learners.
- 59
- 54
- Image Classification Algorithm Based on Improved AlexNet51
The results indicate that the method not only reduces the number of parameters and parameters proportion of the full connection layer, but also improves the classification accuracy compared with AlexNet.
- 50
- 49
- Big graph search: challenges and techniques48
This article introduces the application of graph search in various scenarios, formalizes the graph search problem, and gives an analysis of graphSearch from an evolutionary point of view, followed by the evidences from both the industry and academia.
- 47
- 42
- Distortion-Adaptive Salient Object Detection in 360$^\circ$ Omnidirectional Images37
Results show the proposed baseline approach outperforms the top-performanced state-of-the-art methods on 360 and validate the usefulness of the proposed dataset and approach to boost the development of SOD on 360<inline-formula><tex-math notation="LaTeX">$^\circ$</tex-Math></inline- formula> omnidirectional scenes.
- Fuzzy community detection via modularity guided membership-degree propagation36
A fuzzy community detection method that iteratively propagates membership-degrees of all vertices that is robust on both disjoint and fuzzy community detections and exhibits a high flexibility between computational complexity and overall performance.
- Retinomorphic Object Detection in Asynchronous Visual Streams35
This paper proposes a novel problem setting, retinomorphic object detection, which is the first trial that integrates foveal-like and peripheral-like visual streams and presents a novel bio-inspired unifying framework to fuse two sensing modalities via a dynamic interaction mechanism.
- A Class-Imbalanced Deep Learning Fall Detection Algorithm Using Wearable Sensors35
An algorithm that can effectively distinguish falls from a large amount of ADL signals is proposed, which is effective on class-imbalanced data and more suitable for real-life application compared to previous works.
- 35
- Word categorization from distributional information: Frames confer more than the sum of their (Bigram) parts35
This paper investigated how adults exposed to an artificial language use distributional information to categorize words and found that learners categorized words together when they occur in similar frame contexts, but not when they occurred in similar bigram contexts.
- Classifying objectionable websites based on image content35
The system uses WIPEℳ (Wavelet Image Pornography Elimination) and statistics to provide robust classification of on-line objectionable World Wide Web sites, and has demonstrated 97% sensitivity and 97% specificity in classifying a Web site based solely on images.
- Visual Saliency Computation34
This book covers fundamental principles and computational approaches relevant to visual saliency computation from an innovative perspective that combines both neurobiology and machine learning.
- 33
- 33
- Cross-Reference Stitching Quality Assessment for 360° Omnidirectional Images33
Qualitative and quantitative experiments show the proposed Omnidirectional Stitching Image Quality Assessment (OS-IQA) method outperforms the classic IQA metrics and is highly consistent with human subjective evaluations.
- Learning From Large-Scale Noisy Web Data With Ubiquitous Reweighting for Image Classification33
A Ubiquitous Reweighting Network (URNet) that can learn an image classification model from noisy web data by reweighting the influence of each instance according to different class sizes, large instance clusters, its confidence, small instance bags, and the labels is proposed.
- 33
- 29
- Tagging over time29
It is observed that the addition of this meta-learning layer produces much improved results that outperform best-known results, and the T/T approach produces progressively better annotation with time, significantly outperforming the black-box as well as the static form of the meta-learner, on real-world data.
- Deblurring Neural Radiance Fields with Event-driven Bundle Adjustment28
Event-driven Bundle Adjustment for Deblurring Neural Radiance Fields (EBAD-NeRF) is proposed to jointly optimize the learnable poses and NeRF parameters by leveraging the hybrid event-RGB data to strengthen the explicit modeling of camera motion blur.
- 28
- 28
- 26
- Pairwise comparison and rank learning for image quality assessment26
A rank learning optimization framework is developed to model IQA and the PCs of image quality instead of numerical ratings are incorporated into the optimization framework, and a novel no-reference IQA is proposed to infer image quality in terms of imagequality ranks.
- 25
- Fewer-Shots and Lower-Resolutions25
This paper proposes a few-shot knowledge distillation approach to learn an ultrafast face recognizer via two steps and shows that the proposed approach performs favorably against state-of-the-arts in recognizing low-resolution faces with an extremely low memory of 30KB.
- 24
- Language-Inspired Relation Transfer for Few-Shot Class-Incremental Learning22
A new Language-inspired Relation Transfer (LRT) paradigm to understand objects by joint visual clues and text depictions, composed of two major steps, which outperforms the state-of-the-art models.
- 21
- 21
- Text-prompt Camouflaged Instance Segmentation with Graduated Camouflage Learning20
This work proposes the first Text-Prompt based weakly-supervised camouflaged instance segmentation method named TPNet, leveraging semantic distinctions for effective segmentation, and offers a comprehensive experiment on two common benchmark and demonstrates a significant advancement.
- Rethinking Lightweight Salient Object Detection via Network Depth-Width Tradeoff20
A novel trilateral decoder framework is proposed by decoupling the U-shape structure into three complementary branches, which are devised to confront the dilution of semantic context, loss of spatial structure and absence of boundary detail, respectively.
- View-Aware Salient Object Detection for $360^{\circ }$ Omnidirectional Image20
A view-aware salient object detection method based on a Sample Adaptive View Transformer module with two sub-modules to mitigate issues in panoramic scenarios and outperforms the state-of-the-art methods is proposed.
- 19
- 19
- DanceIt: Music-Inspired Dancing Video Synthesis19
The proposed approach can perform the function of generating promising dancing videos by inputting music and develops a temporal alignment algorithm to align the rhythm of music and dance.
- 18
- 18
- CT image reconstruction on a low dimensional manifold18
This proposed method recovers detailed structural information of images, significantly enhancing spatial and contrast resolution of CT images, and enables a more accurate image reconstruction with high fidelity and Contrast resolution.
- Two-stream Attentive CNNs for Image Retrieval18
Two-stream attentive CNNs for image retrieval has two streams that simultaneously handle two tasks and can be computed as the human-being does by reserving the conspicuous content and suppressing the irrelevant regions.
- 17
- ES2FL: Ensemble Self-Supervised Feature Learning for Small Sample Classification of Hyperspectral Images17
A novel ensemble self-supervised feature-learning method that can automatically learn deep features conducive to classification without any annotation information, significantly reducing the dependence of deep-learning models on massive labeled samples is proposed.
- 17
- Approximating Graph Pattern Queries Using Views17
This paper studies approximation of graph pattern queries using views, and verifies the effectiveness and efficiency of approximating simulation and subgraph query using views.
- Structure‐adaptive Shape Editing for Man‐made Objects17
A framework for real‐time shape editing, which adapts the structure of shape to continuous user editing operations, is proposed, which is capable of both structure‐preserving and structure‐varying shape editing.
- 17
- 16
- 16
- Identifying Emotions Aroused from Paintings16
This work presents an adaptive learning algorithm that leverages labeled photographs and unlabeled paintings to infer the visual appeal of paintings and iteratively adapt the feature distribution in photographs to fit paintings and maximize the joint likelihood of labeled and unl labeled data.
- All-optical UWB signal generation and multicasting using a nonlinear optical loop mirror16
An all-optical scheme for ultra-wideband (UWB) signal generation (positive and negative monocycle and doublet pulses) and multicasting using a nonlinear optical loop mirror (NOLM) is proposed and demonstrated.
- SketchMan15
This work proposes a new challenging task sketch enhancement (SE) defined in an ill-posed space, i.e. enhancing a non-professional sketch (NPS) to a professional sketch (PS), which is a creative generation task different from sketch abstraction, sketch completion and sketch variation.
- 15
- Relaxing Graph Pattern Matching With Explanations15
This study formalizes taxonomy simulation, a natural matching semantics combing graph simulation with taxonomy, and proposes its pattern relaxation to enrich graph pattern matching results with taxonomic information, and designs topological ranking and diversified topologicalranking for top-k relaxations.
- 15
- 15
- 14
- Learning visual saliency from human fixations for stereoscopic images14
A novel stereoscopic saliency detection algorithm by machine learning technique that can predict better saliency results for stereoscopic images than other existing ones is designed.
- 14
- Measuring Visual Surprise Jointly from Intrinsic and Extrinsic Contexts for Image Saliency Estimation14
This work presents an approach to estimate image saliency by measuring the joint visual surprise from intrinsic and extrinsic contexts and adopts a Markov chain to model a surprise-driven attention-shifting process so as to infer the salient patches that can best capture human attention.
- Wavelet frame based color image demosaicing14
Numerical simulations of two datasets: McM and Kodak PhotoCD, show that the proposed wavelet frame based methods outperforms other existing methods in terms of PSNR and visual quality.
- Semantic Contrastive Bootstrapping for Single-Positive Multi-label Recognition13
A semantic contrastive bootstrapping (Scob) approach to gradually recover the cross-object relationships by introducing class activation as semantic guidance and a recurrent semantic masked transformer to extract iconic object-level representations.
- How to Learn a Domain-Adaptive Event Simulator?13
This work proposes one of the first learning-based, domain-adaptive event simulator that learns pixel-wise distributions of event contrast thresholds that, after stochastic sampling and paralleled rendering, can generate event representations well aligned with those from the data from realistic event cameras.
- 12
- Ultrafast Video Attention Prediction with Coupled Knowledge Distillation12
An extremely light-weight network with ultrafast speed, named UVA-Net, constructed based on depth-wise convolutions and takes low-resolution images as input, which is comparable to 11 state-of-the-art models in video attention prediction and costs only 0.68 MB memory footprint.
- Semi-Supervised Cervical Dysplasia Classification with Learnable Graph Convolutional Network12
A novel and more flexible GCN model with a feature encoder that adaptively updates the adjacency matrix during learning and demonstrate that this model design leads to improved performance.
- 12
- 11
- Semantics and aesthetics inference for image search: statistical learning approaches11
This thesis presents algorithms and statistical models for inferring image semantics and aesthetics from visual content, specifically aimed at improving real-world image search and explores the use of image search techniques for designing a novel image-based CAPTCHA, a Web security test aimed at distinguishing humans from machines.
- Two-scale image retrieval with significant meta-information feedback11
A two-scale image retrieval system is developed to provide efficient search in large-scale databases as well as flexibility for users to incorporate ubjective preferences during retrieval to learn about hidden connections between visual patterns and non-imagery types of data.
- 10
- A joint model of extended LDA and IBTM over streaming Chinese short texts10
A joint model based on online algorithms of Latent Dirichlet Allocation (LDA) and BTM is proposed, which combines the merits of both models and alleviates the sparsity when addressing short texts with the online algorithm of BTM, namely Incremental Biterm Topic Model (IBTM).
- 9
- 9
- 8
- Unsupervised Cross-Domain Adaptation for Response Selection Using Self-Supervised and Adversarial Training8
This work proposes a two-stage method to adapt a response selection model to a new domain using self-supervised and adversarial training based on pre-trained language models (PLMs) to efficiently incorporate domain awareness and target-domain knowledge to PLMs.
- 8
- 8
- Collaborative Annotation of Semantic Objects in Images with Multi-granularity Supervisions8
Experimental results show that the proposed collaborative annotation approach can facilitate the annotation and generate object masks that are consistent with those generated by the LabelMe toolbox.
- Penrose high-dynamic-range imaging8
Both quantitative and qualitative results show the superiority of Penrose pixel layout over square pixel layout in HDR imaging, using a Gaussian mixture model for regularization.
- External concept support for group support systems through Web mining8
The experiment results indicate that the concept space mined from the Web contained qualified concepts to stimulate divergent thinking and demonstrate that external concept support in GSS greatly enhanced group productivity for idea generation tasks.
- AI-Driven Self-Evolving Software: A Promising Path Toward Software Automation7
This work introduces AI-Driven Self-Evolving Software, a new form of software that evolves continuously through direct interaction with users that autonomously interprets user requirements, generates and validates code, and integrates new functionalities.
- 7
- Joint self-supervised and reference-guided learning for depth inpainting7
This work proposes a joint self-supervised and reference-guided learning approach for depth inpainting that integrates a convolutional sparse coding module to preserve local contextual information and a hierarchical joint bilateral filter module for filling using specific adjacent information.
- Robust Frame Based X-Ray CT Reconstruction7
A robust wavelet frame regularization based model is proposed for both global reconstruction and interior tomography and outperforms the FBP and SART+TV methods in terms of preservation of sharp edges, mean structural similarity (SSIM), contrast-to-noise ratio, relative error and correlations.
- 7
- Reconfigurable UWB pulse generator based on pulse shaping in a nonlinear optical loop mirror and differential detection7
A reconfigurable impulse radio ultra-wideband pulse generator for various UWB shapes based on a nonlinear optical loop mirror and differential detection and may be suitable for implementation in future low-cost, high-speed, short-range UWB wireless access applications.
- Two‐way Gaussian mixture models for high dimensional classification7
Experiments on several real data sets show that the parsimonious two‐way mixture often outperforms a mixture model without variable grouping; and as a byproduct, significant dimension reduction is achieved.
- Graph neural networks for network science: A mini review6
This letter provides a mini review of different types of GNNs, as well as their applications in important network science tasks, particularly in nodes and edges identification as well as network robustness.
- 6
- Parsing Objects at a Finer Granularity: A Survey6
An in-depth study of the advanced work on fine-grained visual parsing from a new perspective of learning the part relationship, and revisits the universal challenges in fine- grained part segmentation and recognition tasks and proposes new solutions by part relationship learning.
- 6
- 6
- Constructing 3D Model Based on Panoramic Images6
A simple, fast and low cost method to solve the problem of high cost and complicated process, taking panoramic images of the object and finish few steps of simple settings.
- An image feature point matching algorithm based on fixed scale feature transformation6
Simulation results show the feature points extracted by fixed scale transform operator has good stability in case of image translation, rotation, affine transformation and perspective transform and the robustness trend of algorithm good and has not too much useless computation, which can be applied in various image process systems.
- A neural network framework based on ConvNeXt for side‐channel hardware Trojan detection5
A framework that utilizes the continuous wavelet transform to convert time‐series information and employs an improved ConvNeXt network to detect hardware Trojans and significantly improves the accuracy of hardware Trojan detection is proposed.
- Intelligent agent for hyperspectral image classification with noisy labels: a deep reinforcement learning framework5
A deep reinforcement learning method named HyperRefine is proposed to tackle the label noise problem for hyperspectral image classification and can achieve promising performance on artificially added label noise and pseudo-label noise hyperspectral datasets.
- 5
- 5
- Statistical and machine learning methods for immunoprofiling based on single-cell data5
The current state-of-the-art methods for immunoprofiling are discussed, including those for reducing the dimensionality of high-dimensional single-cell data and methods for clustering, classification, and prediction.
- Revisiting Stochastic Learning for Generalizable Person Re-identification5
This paper revisits the stochastic behaviors from two different perspectives and proposes a model that achieves a better generalization performance on unseen target domains without any additional computation costs or auxiliary modules.
- 5
- 5
- 5
- Large-scale image and video search: Challenges, technologies, and trends5
This special issue is intended to bring together the latest research results in all aspects that relate to large-scale image and video search, including high-dimensional indexing, semantic-basedindexing, etc.
- 4
- ForCenNet: Foreground-Centric Network for Document Image Rectification4
This paper introduces Foreground-Centric Network (ForCenNet), a foreground-centric label generation method, which extracts detailed foreground elements from an undistorted image, and introduces a foreground-centric mask mechanism to enhance the distinction between readable and background regions.
- Holistic Correction with Object Prototype for Video Object Segmentation4
A Holistic Correction Network (HCNet) is proposed to adaptively acquire concise object prototypes for holistic correction at semantic, spatial and temporal aspects to robustly and efficiently deal with diverse complex scenarios.
- Exploring Driving-Aware Salient Object Detection via Knowledge Transfer4
A driving task-oriented dataset where pixel-level masks of salient objects have been annotated and it is found that the cross-domain knowledge difference and task-specific scene gap are two main challenges to focus the salient objects when driving.
- 4
- Q-matrix estimation (validation) methods for cognitive diagnosis4
The Q -matrix, which represents important item characteristics by mapping attributes to items has been proved to be the core factor affecting the accuracy of cognitive diagnostic classification and is of great value to study the methods of Q-matrix estimation (validation).
- 4
- 4
- Hierarchical Deep Cosegmentation of Primary Objects in Aerial Videos4
A hierarchical deep cosegmentation approach that repeatedly divides a video into two sub-videos formed by the odd and even frames, respectively is proposed, which remarkably outperforms 17 state-of-the-art methods in segmenting primary objects in various types of aerial videos.
- Image saliency estimation via random walk guided by informativeness and latent signal correlations4
This paper proposes a graph-based approach for image saliency estimation by incorporating both explicit visual cues and latent signal correlations, which achieves impressive performance in extensive experiments.
- Compass Detection Algorithm based on Image Corner4
A novel compass detection algorithm that uses four compasses as the test mask and calculates threshold value based on image intensity feature to improve the Features from Accelerated Segment Test (FAST) algorithm.
- Research on parallel unsupervised classification performance of remote sensing image based on MPI4
The results show that the classification speed of mass amount data parallel remote sensing image unsupervised classification has been greatly improved and the parallel unsuper supervised classification has effect on parallel efficiency.
- 3
- Eliminating Massive Martian Dust Storms from Images of Tianwen-1 via Deep Learning3
Qualitative and quantitative results show that dust storms can be effectively eliminated by the proposed approach, leading to obviously improved topographical and geomorphological details of Mars.
- 3
- 3
- 3
- 3
- Multi-features-based license Plate Detection in Nighttime Environment3
A novel method for license plate detection in nighttime videos with complicated background scenes that has more robust to interference characters and more accurate when compared with other methods is proposed.
- 3
- 3
- Markerless human motion capture by Markov random field and dynamic graph cuts with color constraints3
The algorithm that is used for human motion capture in this paper is based on Markov random field (MRF) and dynamic graph cuts and takes full account of the impact of 3D reconstruction error and integrates humanmotion capture and3D reconstruction into MRF-MAP framework.
- Learning Yourself: Class-Incremental Semantic Segmentation with Language-Inspired Bootstrapped Disentanglement2
A Language-inspired Bootstrapped Disentanglement framework that leverages the prior class semantics of pre-trained visual-language models to guide the model in autonomously disentangling features through Language-guided Prototypical Disentanglement and Manifold Mutual Background Disentanglement.
- 2
- Semantic Segmentation Algorithm of Animal Husbandry Image Based on an Improved U⁃Net Network2
This study offers effective technical support for advancing aquaculture status monitoring and lays a foundation for further development in this field.
- 2
- 2
- Learning Local Feature Descriptor with Motion Attribute For Vision-based Localization2
A fully convolutional network is designed, named MD-Net, to perform motion attribute estimation and feature description simultaneously and can be integrated into a vision-based localization algorithm to improve estimation accuracy significantly.
- 2
- 2
- Structured Prediction in Time Series Data2
This proposal is interested in structured prediction problems in time series data and focuses on prediction task where the outputs are structured and interdependent, contrary to the non-structured prediction which assumes that the Outputs are independent of other predicted outputs.
- Uncovering the Effect of Visual Saliency on Image Retrieval2
Extensive experiments on a generally used image benchmark demonstrate that the new image retrieval system remarkably outperforms the original one and other traditional ones and a novel approach to effectively involve visual saliency into image retrieval systems by a learning process is proposed.
- 2
- 2
- Beyond pixel2
By analyzing recent trends in selecting saliency context, it is found that the usage of adaptive context could be a feasible direction toward better saliency models, which will be explored in the feature research efforts.
- 2
- 2
- 2
- Firm-Generated Online Content in Social Media and Stock Performance: An Event Window Study of Twitter and the S&P 5001
It is revealed that disseminating a greater number of strong image-enhancing FGOC signals, particularly those related to new products and financial matters, significantly enhances stock performance, resulting in higher abnormal stock returns.
- Toward Unified Co-Speech Gesture Generation via Hierarchical Implicit Periodicity Learning1
This paper proposes a unified Hierarchical Implicit Periodicity (HIP) learning approach for audio-inspired 3D gesture generation that outperforms the state-of-the-art co-speech gesture generation methods by both quantitative and qualitative evaluations.
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- Determining the Correct Number of Clusters in the CT Image Segmentation1
A novel validity index which cannot be affected by the shapes of clusters and corrupted by noise of the investigated datasets is defined and this index is helpful in clustering analysis and computer-aided detection system.
- 1
- Attitude/Position Estimation of Monocular Vision Based on Multiple Model Kalman Filter1
A multiple model Kalman filter is proposed for attitude/position estimation of monocular vision, in which the noise disturbance and data loss during camera imaging process is considered.
- 1
- 1
- 1
- Face recognition using composite classifier with 2DPCA1
This paper proposed an approach of face recognition using two-dimensional principal component analysis (2DPCA) and designed a novel composite classifier founded by statistical technique utilizing the advantages of SVM and Logic Regression in field of classification and therefore its accuracy improved a lot.
- Design of embedded intelligent monitoring system based on face recognition1
The experimental results show that the new embedded intelligent monitoring system based on face recognition can work under various light intensities to recognize human face and send alarm information in real time.
- 1
- Anterior cruciate ligament reconstruction model based on anatomical position locating1
An Anatomical Position Locating (APL) approach based on Expectation Maximization (EM) algorithm is proposed and experimental results verify the assumption and demonstrate that the located anatomical position has great serviceability.
- 1
- A novel dynamical 3D smile measurement method to evaluate the effects of face-lifting surgery: Based on the optical structured light strategy1
A novel optical technique to acquire the dynamical data of face muscle with potential to objectively evaluate the effect of surgery is proposed, which shows enough high accuracy of in acquiring the moving information of every point on smile face.
- Multiresolution Non-Local Mean Filter for Infrared Dim and Small Target Background Suppression1
The presented method can suppress complicated background in dim and small target image effectively through visual quality and value index and local gradient statistics is introduced to boost ability of method which suppresses false by background structure.
- –
- –
- Adapting Dense Vision-Language Relationships for Multi-Label Classification With Partial Label–
A Language-driven Dense Semantic Adaptor (LDSA) that excavates prior-adaptive relationships from multimodal pretrained CLIP models and discovers implicit semantic relationships with the prior-adaptive learning scheme is proposed.
- –
- –
- –
- Piecewise data-driven tight frame and $ \ell_0 $ -balanced image denoising method–
The results indicate that the proposed piecewise data-driven tight frame(DDTF) based approach enhances the quality of the reconstructed images, and for images with piecewise textures, it outperforms several representative nonlocal methods in terms of PSNR and the preservation of texture details.
- –
- –
- –
- Towards imbalanced motion: part-decoupling network for video portrait segmentation–
A part-decoupling network (PDNet) for VPS is proposed which unsupervisedly segments portrait into parts and utilizes different attentiveness on discriminative features specified to each different part, so that the portraits can be segmented more accurately.
- –
- –
- Investigation of Traffic Classification Applied to an Astronomical Data Transmission Network of the XAO Using Deep Learning–
A classification method applied to types of traffic containing astronomical data using deep learning using a convolutional neural network model is proposed, and experimental results indicate that the proposed method can accurately classify types of Traffic containing Astronomical data.
- –
- –
- Complementary Segmentation of Primary Video Objects with Reversible Flows–
This work designs a complementary segmentation network (CSNet) with two separate branches, which can simultaneously encode the foreground and background information along with joint spatial constraints, and designs a neighborhood reversible flow that reflects the most reliable temporal correspondences between superpixels in far-away frames.
- –
- Discussion on The Data Extraction Strategy Of FAQ System Based On Chat Records–
A FAQ system using the optimal data extraction strategy of question answering from machine learning and non-machine learning is developed, the system results show that the performance is reliable.
- –
- Attention Enhanced Multi-patch Deformable Network for Image Deblurring–
An enhanced network based on the four-layer multi-path hierarchy structure which divides image into multi-patches instead of down-sampling or other lossy way is proposed, which has improved nearly 0.4dB PSNR on the quantitative performance on GoPro testing data set, and also achieved clearer visual effect on several scenes.
- –
- Research on Dynamical Control Algorithms Rapid Validation Platform for Mobile Robots–
A realization method of semi-physical real-time simulation platform for mobile robot based on Quanser system can be used for fast verification of the complex control algorithm of mobile robot, and the effectiveness of the method is verified by the trajectory tracking control experiment of the semi- physical simulation system.
- –
- Infrared Image Enhancement Based on Multiscale Bilateral Detail Decomposition–
Experimental results show that the proposed MBDD method outperforms other conventional methods for infrared image enhancement.
- Light-weight binary code embedding of local feature distribution in image search–
Experimental results on several public benchmarks show that, with the proposed two embedding schemes, image search achieves comparable performance with state-of-the-arts with much lower computational cost and memory usage.
- –
- Mining Cluster-Specific Knowledge for Saliency Ranking–
This chapter proposes to model the problem of visual saliency computation in a learning to rank framework, which is proved to be more effective than the classification and regression framework.
- Removing Label Ambiguity in Training Saliency Model–
This Chapter discusses the label ambiguities in training visual saliency models and adopts a multi-instance learning to rank framework to iteratively recover the correct sample labels and optimize the saliency model.
- –
- –
- Notice of Retraction: A Method for Obtaining Regions of Interest with Adaptive Gaussian Mixture Model–
From the view of the tracking effect, it can be concluded that this method of removal shadow of regions of interest can improve the precision rate of segment of moving people and is good for subsequent tracking.
- A Design Framework for Online Discus-sion Speech Act Analysis and Situation Profiling–
This research developed a design framework for systems supporting online discussion speech act analysis and situation profiling that provides guidelines for the creation of speech act category, the choice of appropriate classification algorithms, features, and feature selection techniques necessary to effectively identify the roles each message plays.
- –
- Color correction of the video image obtained by angioscope–
A simple color correction method for the color reproduction of image obtained by angioscope is reported and two matrices are suggested: an average color correction matrix A and a blood vessel color Correction matrix A'.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-14. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.