Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks276 from public data
- Cross-Modal Self-Attention Network for Referring Image Segmentation603
A cross-modal self-attention (CMSA) module that effectively captures the long-range dependencies between linguistic and visual features and a gated multi-level fusion module to selectively integrateSelf-attentive cross- modal features corresponding to different levels in the image.
- 493
- Multiview Spectral Clustering via Structured Low-Rank Matrix Factorization389
An iterative multiview agreement strategy by minimizing the divergence objective among all factorized latent data-cluster representations during each iteration of optimization process, where such latent representation from each view serves to regulate those from other views, and such an intuitive process iteratively coordinates all views to be agreeable.
- Occlusion Aware Unsupervised Learning of Optical Flow344
This work introduces a new method which models occlusion explicitly and a new warping way that facilitates the learning of large motion and shows promising results on Flying Chairs, MPI-Sintel and KITTI benchmark datasets.
- 323
- 283
- A Discriminative Latent Model of Object Classes and Attributes256
This work presents a discriminatively trained model for joint modelling of object class labels and their visual attributes and captures the correlations among attributes using an undirected graphical model built from training data.
- 254
- Survey on Deep Multi-modal Data Analytics: Collaboration, Rivalry, and Fusion250
A substantial overview of the existing state-of-the-arts in the field of multi-modal data analytics from shallow to deep spaces is provided and it is indicated that the critical components for this field go to collaboration, adversarial competition, and fusion over multi- modal spaces.
- 234
- Where-and-When to Look: Deep Siamese Attention Networks for Video-Based Person Re-Identification231
The proposed Siamese model is end-to-end trainable to jointly learn comparable hidden representations for paired pedestrian videos and their similarity value and outperforming state-of-the-art methods.
- Keyphrase Extraction Using Deep Recurrent Neural Networks on Twitter198
A novel deep recurrent neural network model is proposed to combine keywords and context information to perform keyphrases from tweets and the experimental results showed that the proposed method performs significantly better than previous methods.
- Deep adaptive feature embedding with local sample distributions for person re-identification170
A novel objective function is proposed to jointly optimize similarity metric learning, local positive mining and robust deep feature embedding for person re-id by proposing a novel sampling to mine suitable positives within a local range to improve the deep embedding in the context of large intra-class variations.
- High Resolution Acquisition, Learning and Transfer of Dynamic 3‐D Facial Expressions165
New expressions can be synthesized, either as dynamic morphing between individuals, or as expression transfer from a source face to a target face, as demonstrated in a series of experiments.
- Beyond Actions: Discriminative Models for Contextual Group Activities163
The proposed model jointly captures the group activity, the individual person actions, and the interactions among them, and implicitly infer it during learning and inference can significantly improve activity recognition performance.
- Effective Multi-Query Expansions: Collaborative Deep Networks for Robust Landmark Retrieval159
This paper proposes a novel framework, namely, multi-query expansions, to retrieve semantically robust landmarks by two steps, and proposes to learn a collaborative deep networks-based semantically, nonlinear, and high-level features over the latent factor for landmark photo as the training set.
- Manifold-Based Incomplete Multi-View Clustering via Bi-Consistency Guidance135
A novel method termed Manifold-based Incomplete Multi-view clustering via Bi-consistency guidance (MIMB), which flexibly recovers incomplete data among various views, and attempts to achieve biconsistency guidance via reverse regularization.
- What-and-where to match: Deep spatially multiplicative integration networks for person re-identification129
A novel deep multiplicative integration gating function is proposed, which answers the question of what-and-where to match for effective person re-id and is designed to be end-to-end trainable to characterize local pairwise feature interactions in a spatially aligned manner.
- Try This Instead127
This paper proposes attribute-aware collaborative filtering (A2CF) to perform substitute recommendation by addressing issues from both personalization and interpretability perspectives, and proves the recommendation effectiveness and interpretation quality of A2CF.
- Beyond Low-Rank Representations: Orthogonal clustering basis reconstruction with optimized graph structure for multi-view spectral clustering121
This paper decomposes LRR into latent clustered orthogonal representation via low-rank matrix factorization, to encode the more flexible cluster structures than LRR over primal data objects and converts the problem of L RR into that of simultaneously learning Orthogonal clustered representation and optimized local graph structure for each view.
- Every Pixel Counts: Unsupervised Geometry Learning with Holistic 3D Motion Understanding115
Experiments on KITTI 2015 dataset show that the estimated geometry, 3D motion and moving object masks, not only are constrained to be consistent, but also significantly outperforms other SOTA algorithms, demonstrating the benefits of the approach.
- 110
- 3-D PersonVLAD: Learning Deep Global Representations for Video-Based Person Reidentification109
This paper presents the global deep video representation learning to video-based person reidentification (re-ID) that aggregates local 3-D features across the entire video extent to be complementary to3-D ConvNets as a novel layer to capture the appearance and motion dynamics in full-length videos.
- Few-Shot Deep Adversarial Learning for Video-Based Person Re-Identification92
A novel few-shot deep learning approach to video-based person re-identification, to learn comparable representations that are discriminative and view-invariant in matching persons.
- High-order pattern discovery from discrete-valued data91
The authors propose a novel method to discover qualitative and quantitative patterns (or event associations) inherent in a data set that uses the adjusted residual analysis in statistics to test the significance of the occurrence of a pattern candidate against its expectation.
- LBMCH88
A novel method to learning bridging mapping for cross-modal hashing, named LBMCH, is proposed to characterize the cross- modal semantic correspondence by seamlessly connecting these distinct hamming spaces with each preserving the local structure of data objects from an individual modality.
- Cross-Entropy Adversarial View Adaptation for Person Re-identification86
This work learns view-invariant subspace for person re-ID, and its corresponding similarity metric using an adversarial view adaptation approach, and achieves notably improved performance in comparison with the state-of-the-arts on benchmark datasets.
- 86
- 81
- 2D/3D Multimode Medical Image Alignment Based on Spatial Histograms80
A two-dimensional weighted spatial histogram of gradient directions is used to extract statistical features, overcome the algorithm’s limitations, and expand the applicable scenarios under the premise of ensuring accuracy.
- 72
- 72
- 71
- Structured deep hashing with convolutional neural networks for fast person re-identification67
A novel deep hashing framework with Convolutional Neural Networks (CNNs) for fast person re-identification that simultaneously learns both CNN features and hash functions to get robust yet discriminative features and similarity-preserving hash codes.
- DECORE: Deep Compression with Reinforcement Learning64
DECORE assigns an agent to each channel in the network along with a light policy gradient method to learn which neurons or channels to be kept or removed, which leads to a much faster training process compared to existing approaches.
- Thinking inside The Box64
The hypercube recommender (CubeRec) is designed to adaptively learn group hypercubes from user embeddings with minimal information loss during preference aggregation, and to leverage a revamped distance metric to measure the affinity between grouphypercubes and item points.
- Multi-Graph Heterogeneous Interaction Fusion for Social Recommendation60
This article proposes a novel social recommendation approach, namely Multi-Graph Heterogeneous Interaction Fusion (MG-HIF), to fuse heterogeneous interaction features from multi-graphs, i.e., user–item bipartite graph and social relation network to improve the vertex representation learning.
- 59
- Adaptive Data-Free Quantization58
An Adaptive Data-Free Quantization (AdaDFQ) method is proposed, which revisits DFQ from a zero-sum game perspective upon the sample adaptability between two players — a generator and a quantized network, and reveals the largest adaptability is NOT the best for sample generation to benefit Q's generalization.
- 57
- Design and Research of Intelligent Question-Answering(Q&A) System Based on High School Course Knowledge Graph57
The system, which is the first innovative application in the field of high school teaching, integrates knowledge graph technology and intelligent question answering technology, and introduces big data technology.
- Twin-Incoherent Self-Expressive Locality-Adaptive Latent Dictionary Pair Learning for Classification56
- 53
- 52
- Exploiting Correlation Consensus49
This paper proposes a novel angular based regularizer coupled with the authors' objective function, which is aided by trace lasso and minimized to yield sparse representation vectors encoding data correlations in multiple modalities, which can generate a compatible data similarity matrix for multi-modal data.
- 47
- Dense Residual Network: Enhancing global dense feature flow for character recognition46
This paper proposes an effective convolutional deep model termed Dense Residual Network (DRN) for the task of optical character recognition and proposes a refined residual dense block (r-RDB) to retain the ability of local feature fusion and local residual learning of original RDB to enhance the local and global feature learning abilities of DenseNet.
- Robust hashing for multi-view data: Jointly learning low-rank kernelized similarity consensus and hash functions46
This paper motivates the problem of jointly and efficiently training the robust hash functions over data objects with multi-feature representations which may be noise corrupted, and proposes an approach to effectively and efficiently learning low-rank kernelized similarity consensus and hash functions.
- 46
- Diffusion-Based Cloud-Edge-Device Collaborative Learning for Next POI Recommendations45
A novel collaborative learning framework, Diffusion-Based Cloud-Edge-Device Collaborative Learning for Next POI Recommendations (DCPR), leveraging the diffusion model known for its success across various domains to offer region-specific and highly personalized POI recommendations while reducing on-device computational burdens.
- 44
- Few-Shot Scene Adaptive Crowd Counting Using Meta-Learning43
The proposed approach learns the model parameters in a way that facilitates the fast adaptation to the target scene with a few gradient updates to the learned parameters and outperforms other alternatives in few-shot scene adaptive crowd counting.
- 41
- Fully Point-wise Convolutional Neural Network for Modeling Statistical Regularities in Natural Images38
This paper proposes to randomly shuffle the pixels in the origin images and leverage the shuffled image as input to make CNN more concerned with the statistical properties and shows the superiority of the proposed network over the existing architectures, i.e., using 1/10~1/100 network parameters and computational cost while achieving comparable performance.
- Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification37
An alternative SR-free paradigm to directly compare HR and LR images via a dynamic metric that is adaptive to the resolution of a query image is proposed and achieves state-of-the-art performance on multiple CRReID benchmarks.
- 37
- Learning Hybrid Representation by Robust Dictionary Learning in Factorized Compressed Space37
The robust dictionary learning (DL) is investigated to discover the hybrid salient low-rank and sparse representation in a factorized compressed space and a discriminative J-RFDL model is presented, which can improve the discriminating abilities of learnt coefficients by minimizing the classification error jointly.
- 37
- Boosting Image De-Raining via Central-Surrounding Synergistic Convolution36
Evaluations of twelve de-raining methods on nine public datasets demonstrate that the proposed Synergistic Convolution can comprehensively improve the performance of twelve de-raining networks under various rainy conditions without changing the original network structure or introducing extra computational costs.
- 36
- Bayesian Personalized Feature Interaction Selection for Factorization Machines36
A Bayesian Personalized Feature Interaction Selection (BP-FIS) mechanism under the Bayesian Variable Selection (BVS) theory is proposed by proposing a Bayesian generative model and derived the Evidence Lower Bound (ELBO), which can be optimized by an efficient Stochastic Gradient Variational Bayes method to learn the parameters.
- Bayesian feature interaction selection for factorization machines34
Novel Bayesian variable selection methods, targeting feature interaction selection for factorization machines, which effectively reduce the number of interactions are proposed.
- Shifting Hypergraphs by Probabilistic Voting34
This paper develops a novel paradigm, namely hypergraph shift, to find robust graph modes by probabilistic voting strategy, which are semantically sound besides the self-cohesiveness requirement in forming graph modes, and formally proves that the technique is always convergent.
- Learning Hybrid Representations for Automatic 3D Vessel Centerline Extraction33
This work proposes a hybrid representation learning approach to address the challenge of preserving the continuity of extracted vessels when segmenting such thin tubular structures from 3D images, and validates the proposed approach on CTA datasets and demonstrates its superior performance compared to both traditional and CNN-based baselines.
- Crossing generative adversarial networks for cross-view person re-identification33
The proposed model consists of the variational auto-encoder to encode the pair into respective latent variables, a proposed cross-view alignment to reduce the view disparity, and an adversarial layer to seek the joint distribution of latent representations.
- Clustering via geometric median shift over Riemannian manifolds33
This work presents a novel approach for the median shift over Riemannian manifolds that uses the geometric median and geodesic distances and demonstrates that this method outperforms the state-of-the-art algorithms in data clustering, image segmentation, and noise filtering.
- 28
- 28
- Saliency detection via multi-level integration and multi-scale fusion neural networks28
An end-to-end multi-level feature integration and multi-scale feature fusion network to better predict salient objects in challenging images and outperforms the state-of-the-art saliency models on several challenging image datasets.
- Local Variational Feature-Based Similarity Models for Recommending Top- N New Items27
A Bayesian generative model is proposed, called the Local Variational Feature-based Similarity Model, to encapsulate local and global similarity functions, and a variational Expectation Minimization algorithm is presented for efficient approximate inference.
- Graph-based Text Classification by Contrastive Learning with Text-level Graph Augmentation26
A novel deep TC model, namely Text-level Graph Networks with Contrastive Learning (TGNcl), is developed and the empirical results demonstrate that TGNcl can outperform the existing state-of-the-art TC models.
- MAPD: An improved multi-attribute pedestrian detection in a crowd26
An improved Multi-attribute pedestrian detection (MAPD) method, which coptimize person intra-class compactness and inter-class discrepancy, and proposes a novel Piecewise NMS algorithm to reduce false positive of small objects.
- Recovering Accurate Labeling Information from Partially Valid Data for Effective Multi-Label Learning26
A novel two-stage PML method, namely Partial Multi-Label Learning with Label Enrichment-Recovery (PML3ER), where in the first stage, it estimates the label enrichment with unconstrained label propagation, then jointly learns the ground-truth confidence and multi-label predictor given the label enriched.
- An Iterative Fusion Approach to Graph-Based Semi-Supervised Learning from Multiple Views25
An iterative Semi-Supervised Metric Fusion (SSMF) approach that fuses metrics and label propagation results from multi-views iteratively until the fused metric and label propagate results converge simultaneously is developed.
- 24
- Block-Aware Item Similarity Models for Top- N Recommendation22
By regularizing the item similarity matrix of item similarity models with BDR, this work obtains a block-aware item similarity model and shows that the block-diagonal structure is crucial to the performance of top-N recommendation.
- 22
- M-AR: A Visual Representation of Manual Operation Precision in AR Assembly21
A high-precision cognitive AR representation (M-AR) for manual operation guidance is proposed, which aims to transform a high- Precision assembly relationship into a series of visual information to meet the user’s cognition of high-Precision task intention.
- Lightweight Embeddings for Graph Collaborative Filtering20
This work proposes Lightweight Embeddings for Graph Collaborative Filtering (LEGCF), a parameter-efficient embedding framework dedicated to GNN-based recommenders, which innovatively introduces an assignment matrix as an additional learnable component on top of meta-embeddings.
- 20
- 20
- 19
- Multi-view laplacian least squares for human emotion recognition19
To meet the requirements of large-scale applications, weighted local preserving embedding is proposed as the out-of-sample extension of MvLLS, basing on the idea of maintaining the manifold structures of original space.
- 19
- 18
- 18
- Small-target ship detection in SAR images based on densely connected deep neural network with attention in complex scenes18
A model based on a densely connected deep neural network with an attention mechanism that achieves an optimal performance between the accuracy and calculation time compared to other state-of-the-art detection methods on public synthetic aperture radar ship detection datasets is proposed.
- A Robust Approach for Scene Text Detection and Tracking in Video18
An effective deep neural network model for detecting text in individual video frames is proposed, which enhances the EAST model by introducing deconvolution layers and inception modules, and a correlation filter based tracking algorithm for text in the video is presented.
- Shifting multi-hypergraphs via collaborative probabilistic voting18
This paper proposes a novel paradigm aimed at shifting high-order edges (i.e., hyperedges) to deliver graph modes, via a novel probabilistic voting strategy, named SMHCPV, so as to further improve the performance over hypergraph shift method.
- Emotion-aware Personalized Music Recommendation with a Heterogeneity-aware Deep Bayesian Network17
A Heterogeneity-aware Deep Bayesian Network (HDBN) is proposed to model four types of heterogeneity that an EMRS should account for: emotion heterogeneity across users, emotion heterogeneity within a user, music mood preference heterogeneity across users, and music mood preference heterogeneity within a user.
- 17
- 17
- Deep Adversarial Inconsistent Cognitive Sampling for Multiview Progressive Subspace Clustering16
A novel deep adversarial inconsistent cognitive sampling (DAICS) method for multiview progressive subspace clustering and demonstrates the superiority of DAICS over state-of-the-art methods.
- 16
- 16
- 16
- Optimal parameters of the SVM for temperature prediction16
The developed genetic algorithm optimization parameters of the predicted values were closest to the measured value after the analog trend, and it is the most fitting measured value trends, and its homing speed is relatively fast.
- Detecting image forgeries using metrology16
This paper presents a framework for detecting fake regions using single view metrology and enforcing geometric constraints from shadows and shows that image forgery on the vertical plane or arbitrary plane can be detected through the measurement on such plane.
- Self‐supervised representation learning of metro interior noise based on variational autoencoder and deep embedding clustering15
The experimental results show that, compared to traditional mechanism‐based models for characterizing interior noise, this approach offers a data‐driven general analysis framework, providing a foundational model for downstream tasks.
- 15
- Quaternion Factorization Machines: A Lightweight Solution to Intricate Feature Interaction Modeling15
By introducing a brand new take on FM-based models with the notion of quaternion algebra, these models not only enable expressive inter-component feature interactions but also significantly reduce the parameter size due to lower degrees of freedom in the hypercomplex Hamilton product compared with real-valued matrix multiplication.
- Pyramid fully residual network for single image de-raining15
Comprehensive experimental results show that the PFR-Net outperforms the state-of-the-art single de-raining methods with a big margin on Rain100H, Rain100L and Rain1200 datasets.
- Three-Stage Bidirectional Interaction Network for Efficient RGB-D Salient Object Detection14
This work remodel the relationship between RGB features and depth features from a new perspective of the feature encoding stage and proposes a threestage bidirectional interaction network (TBINet), which achieves the best performance and efficiency trade-off.
- 14
- An Improved YOLOv7 Model Based on Visual Attention Fusion: Application to the Recognition of Bouncing Locks in Substation Power Cabinets14
An improved GF-YOLOv7 network model is proposed to apply in the recognition of the status of bounce locks in a substation, and the accuracy reaches 97.8%, which can meet the needs of the detection task of substation bounce locks.
- Diversifying Inference Path Selection: Moving-Mobile-Network for Landmark Recognition14
This paper proposes a novel Moving-Mobile-Network, named M2Net, for landmark recognition, equipped each landmark image with located geographic information, and intuitively finds that M2 net can essentially promote the diversity of the inference path (selected blocks subset) selection, so as to enhance the recognition accuracy.
- An efficient framework of Bregman divergence optimization for co-ranking images and tags in a heterogeneous network14
This paper presents a novel co-ranking method with random walks that is able to significantly improve the ranking effectiveness on both images and tags and improves the performance of the algorithm in computational complexity and the out-of-sample problem.
- Detecting Image Forgeries using Geometric Cues13
This chapter presents a framework for detecting fake regions by using various methods including watermarking technique and blind approaches, and describes current categories on blind approaches which can be divided into five: pixel-based techniques, format- based techniques, camera-basedtechniques, physically-based Techniques and geometric-based technique.
- 12
- Topic extraction from extremely short texts with variational manifold regularization12
This work extends Dirichlet multinomial mixture to a novel Laplacian Dirichlet Multinomial Mixture topic model for short texts, and develops an online version of LapDMM, namely Online LapDMM, to achieve inference speedup on massive short texts.
- Adaptive Structure-Constrained Robust Latent Low-Rank Coding for Image Recovery12
A robust representation learning model called Adaptive Structure-constrained Low-Rank Coding (AS-LRC) for the latent representation of data and selects the L2, 1-norm on the projection for extracting group sparse features rather than learning low-rank features by Nuclear-norm regularization, which can make learnt features robust to noise and outliers in samples.
- Face Verification with Multi-Task and Multi-Scale Feature Fusion12
This work explores using identification signals to supervise one CNN and the combination of semi-verification and identification to train the other one, and proposes a method based on two deep convolutional neural networks for face verification.
- 12
- 11
- 11
- Video Text Detection by Attentive Spatiotemporal Fusion of Deep Convolutional Features11
A novel end-to-end deep neural network for detecting scene text in the video, which combines complementary text features from multiple related frames to enhance the overall detection performance relative to single-frame detection schemes.
- 11
- 10
- Reasoning Under Uncertainty: Exploring Probabilistic Reasoning Capabilities of LLMs9
This work presents the first comprehensive study of the reasoning capabilities of LLMs over explicit discrete probability distributions, evaluating models on three carefully designed tasks, mode identification, maximum likelihood estimation, and sample generation.
- Few-Shot Referring Video Single- and Multi-Object Segmentation Via Cross-Modal Affinity with Instance Sequence Matching9
A newly designed cross-modal affinity (CMA) module based on a Transformer architecture facilitates the establishment of multi-modal affinity over a limited number of samples, allowing the rapid acquisition of new semantic information while fostering the model’s adaptability to diverse scenarios.
- 9
- Adaptive Multi-Task Dual-Structured Learning with Its Application on Alzheimer’s Disease Study9
This research presents a novel approach called "Smart Machines" that automates the very labor-intensive and therefore time-heavy and expensive and expensive process of manually rating the disease severity and predicting corresponding clinical symptoms through multi-task learning.
- 9
- A deep learning-based algorithm for the detection of personal protective equipment8
The improved model incorporates rich multi-level and multi-scale features, allowing it to better adapt to tasks involving geometric transformations and aligning with the image data distribution in construction scenarios.
- High performance dilated convolutions on multi-core DSPs8
This paper introduces a high-performance parallel direct implementation of dilated convolutions on multi-core DSPs in a CPU-DSP heterogeneous prototype processor, which can effectively capture the data locality in dilated convolutions.
- 8
- 8
- Weakly labeled fine-grained classification with hierarchy relationship of fine and coarse labels8
This paper proposes to make fine labels prediction under a weakly supervised setting where a subset of training data is labeled with fine labels and the others only have coarse labels, and aims to explore the hierarchy relationship between coarse classes and fine classes to achieve a better performance on fine-grained classification.
- 8
- Deep Learning Models for Predicting Phenotypic Traits and Diseases from Omics Data8
It is shown that integration of multi-omics profiles into DNN-based learning methods could improve the prediction of the molecular subtypes of breast cancer and the superiority of the proposed DNN models over the SVM model for predicting triglyceride concentrations is demonstrated.
- Deep Co-attention based Comparators For Relative Representation Learning in Person Re-identification8
- 7
- 7
- Forecasting the Dst Index with Temporal Convolutional Network and Integrated Gradients7
A Temporal Convolutional Network is deployed in tandem with the Integrated Gradient algorithm to predict the Dst index and scrutinize its associated physical processes, affirm that the TCN+IG technique not only enhances space-weather forecast accuracy but also advances the comprehension of the underlying physical processes in space weather.
- A non-negative feedback self-distillation method for salient object detection7
A non-negative feedback loss is proposed for SOD, which uses different ways to calculate the distillation loss of the foreground and background respectively, to ensure that the teacher network transfers only positive knowledge to the student.
- 7
- 7
- 7
- 7
- 7
- 7
- Learning Transferable Distance Functions for Human Action Recognition7
This chapter considers the problem of human action recognition from a single clip per action, and shows that the transferable distance function learning not only improves the recognition accuracy of the single clip action recognition, but also significantly enhances the efficiency of the matching scheme.
- 6
- 6
- 6
- 6
- Perceptual loss function for generating high-resolution climate data6
This paper studies the perceptual loss when increasing the resolution of raw precipitation data by ×4 and ×8 under training modes of CNN and GAN and proposes a new approach to benefit from perceptual loss while considering the characteristics of climate data.
- 6
- Partial-Label and Structure-constrained Deep Coupled Factorization Network6
This paper technically proposes an enriched prior guided framework, called Dual-constrained Deep Semi-Supervised Coupled Factorization Network (DS2CF-Net), for discovering hierarchical coupled data representation, which designs a deep factorization architecture using multilayers of linear transformations, which can coupled update both the basis vectors and new representations in each layer.
- 6
- 6
- Convolutional Temporal Attention Model for Video-Based Person Re-Identification6
A temporal attention approach for aggregating frame-level features into a video-level feature vector for re-identification and a fully convolutional temporal attention model for generating the attention scores is proposed.
- 6
- 6
- 6
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- Research on face detection method based on improved MTCNN network5
A face detection algorithm based on improved Multi-Task Convolution Neural Network (MTCNN) is proposed, which improves the network structure of MTCNN, builds a neural network model based on M TCNN using TensorFlow, and cascades an eye regression network.
- Approximate Optimal Transport for Continuous Densities with Copulas5
Empirical results demonstrate that the novel Cop-OT can gain more accurate approximations to continuous OT values than the state-of-the-art baselines and can be efficiently solved by stochastic optimization with less time and memory requirements.
- ABSTRACT: Cloud-based Image Classification Service is Not Robust to Affine Transformation: A Forgotten Battlefield5
This paper makes the first attempt to conduct an extensive empirical study of Affine Transformation (AT) attacks against main stream real-world cloud-based classification services and proposes two defense algorithms to address these security challenges.
- 5
- Person Re-Identification by Localizing Discriminative Regions5
This paper introduces an end-to-end Siamese convolutional neural network that firstly localizes discriminative salient image regions and then computes the similarity based on these image regions in conjunction with the whole image.
- Design and Realization of the 3D Electronic Chart Based on GIS and Virtual Reality Technology5
A 3D electronic chart system framework based on GIS and virtual reality technology and studied the key technologies has been developed for Sutong Bridge section of Yangtze River waterway and it is verified to be feasible and reliable for ensuring safety of navigation.
- Research on Scoreboard Detection and Localization in Basketball Video5
An algorithm of detection and location for scoreboard is proposed using multi-dimensional correlation and the template of scoreboard is built using multi-dimensional correlation, it involves the location and the detection parameter.
- A comparison of the performance of Chinese large language models and ChatGPT throughout the entire clinical workflow4
Evaluating the capacity of the Chinese large language models for providing continuous clinical decision support by assessing their performance with simulated patient cases found LLMs outperform human emergency fellow physicians and exhibit rapid development, offering significant practical application potential in healthcare.
- Advances in the authentication of collagen products based on DNA technology4
The development of a complete set of DNA authentication is of great significance for the control of collagen products quality and will contribute to sustainable development of collagen industry.
- Non-local self-similarity recurrent neural network: dataset and study4
By measuring the self-similarity of the neighborhood features, NSRNN model outperforms other state-of-the-art methods in term of image denoising performance.
- Real-time instance segmentation with assembly parallel task4
This paper proposes a real-time framework denoted by APTMask for instance segmentation, which builds on the real-time project YOLACT and devise a new mask branch, which can more effectively exploit the semantic information of PA-FPN deeper features and the positional information of shallow features for mask representation.
- IMCI: an efficient fingerprint retrieval approach based on 3D stacked memory4
Data deduplication as a type of redundant data elimination technology can effectively reduce the impact of redundancy on storage costs, which consequently alleviates the problems and pressures caused by massive data storage, management, and backup.
- Homoglyph attack detection with unpaired data4
This paper introduces a new unpaired homoglyph attack detection system using a convolutional neural network and forms two unpaired datasets based on the original datasets reported in [36], which contain real and spoof names for both domains and processes.
- 4
- ChipDiff: Staged diffusion model with loss gradient guidance for Chinese ink painting style transfer3
- 3
- 3
- 3
- Attention Based Spatial-Temporal Graph Convolutional Networks for RSU Communication Load Forecasting3
- 3
- Deep Learning in Person Re-identification for Cyber-Physical Surveillance Systems3
The Cyber-physical Systems are a combination of integrated physical processes, networking and computation to be minored and controlled y embedded subsystems via networked systems with feedback loops to change their behaviour when needed.
- 3
- 3
- 3
- Human Action Recognition from Video Sequences by Enforcing Tri-view Constraints3
A new view-invariant measure for human action recognition by enforcing tri-view constraints is proposed that can identify human pose transitions, in spite of dynamic time-lines, different viewpoints and unknown camera parameters.
- 3
- 2
- Screening and identification of versican as a sensitive biomarker and potential therapeutic target in basal cell carcinoma2
The cell-derived xenograft model study in nude mice showed that knockdown of VCAN in tumour cells significantly suppressed tumour size compared to control tumour cells, suggesting that VCAN is one of the important genes for tumourigenesis.
- Self-Knowledge Distillation For the Object Segmentation Based on Atrous Spatial Pyramid2
An effective Self-Knowledge Distillation framework via Atrous Spatial Pyramid Structure (ASPS) is presented, which is able to enhance the performance of the object segmentation network without increasing network parameters.
- WA-Transformer: Window Attention-based Transformer with Two-stage Strategy for Multi-task Audio Source Separation2
The proposed WA-Transformer adopts self-attention and window attention blocks to model global dependencies and local correlation in a parameter-efficient way for multi-task audio source separation and has advantages over several well-known models.
- 2
- 2
- 2
- 2
- 2
- Optimization based on nonlinear transformation in decision space2
A new framework called optimization based on nonlinear transformation in decision space (ONTD) is proposed, in which a problem population is generated by converting a given problem, and each converted problem (subproblem) has its own interesting area with high calculation weight on the decision space.
- 2
- 2
- Research on Uncertain Network Design Problem2
A bi-level and stochastic programming model for the discrete network design problem is proposed and the results show that the Stochastic model is more flexible and practical compared with the deterministic one.
- 2
- Single-Sample Face Recognition Using SIFT Features2
The pattern-specific SIFT feature and a simple non-statistical matching strategy combined with feature clustering to solve single training sample face recognition problems with only one training sample.
- Computer-aided characterization of solitary pulmonary nodules (SPNs) using structural 3D, texture, and functional dynamic contrast features2
Combining 3D structural, textural and functional contrast features can provide a more comprehensive examination of the SPNs by coupling dynamic CT scan techniques with image processing to quantify multiple properties that relate to tumor geometry and tumor angiogenesis.
- Automated endometrial segmentation, thickness measurement and pattern prediction on uterine ultrasound images1
A novel automated method for routine endometrial thickness and pattern assessment is demonstrated and is reported to be a first-of-its kind method for automated endometrial pattern assignment (homogenous / triple-line) on uterine ultrasound images.
- 1
- 1
- 1
- 1
- Enhanced Feature Extraction with AL-YOLOv9s Lightweight Model: Application in Key Component Recognition Within Highly Integrated Device Environments1
The AL-YOLOv9s model is based on the advanced YOLOv9s algorithm and incorporates an enhanced efficient multi-scale attention module to boost feature extraction capabilities and improves the efficiency and accuracy of detecting circuit breaker lockouts.
- A Multi-Model Fusion Framework to Enhance sEMG-Based Gesture Recognition Assisted by Visual Features1
- Image-driven Audio-visual Universal Source Separation1
An image-driven audio-visual universal source separation (ID-USS) and the proposed ID-USS-Conformer, which has achieved a 10.139 dB signal-to-distortion ratio improvement in the test set and outperformed the compared methods.
- 1
- 1
- Research on the Application of YOLOv3-Tiny Algorithm in the Detection of Circuit Experiment Equipment1
A real-time target detection model based on YOLOv3-Tiny is proposed for electronic equipment category imbalance by data augmentation of self-made circuit experimental equipment data set, and the problem of electronic equipment category imbalance is solved.
- Research on Wireless Video Monitoring System Based on Hisilicon Platform1
A wireless intelligent video surveillance system based on the Hisilicon platform using the H.264 encoding protocol to compress and encode video streams and implementing the MOSSE algorithm to track and detect the moving objects in the video.
- 1
- 1
- Different Feature Selection of Soil Attributes Influenced Clustering Performance on Soil Datasets1
The results demonstrated that combination of spectral clustering algorithm with the feature subsets containing environmental attributes rather than spatial attributes may be a better choice in applications of soil data clustering.
- 1
- 1
- 1
- 1
- 1
- Deep CNNs for Face Verification1
A method based on two deep convolutional neural networks for face verification using different landmarks of faces to solve the problems caused by poses and achieves 99.71 % verification accuracy on LFW dataset.
- Robust clustering for social annotated images1
A dynamic and efficient thresholding scheme for adaptively terminate the spectral clustering process and a new image clustering method using semantic labels and their co-occurrence statistics and designs the similarity metric for social annotated images are proposed.
- 1
- 1
- –
- Enhancing part-level text-to-shape generation through hierarchical structure modeling–
HierT2S, a hierarchical structure modeling framework that integrates a hierarchical text tree representation with a conditional diffusion model for part-level text-to-shape generation, is introduced, offering a structure-aware approach to text-to-3D generation.
- –
- –
- –
- Cross-attention Secretly Performs Orthogonal Alignment in Recommendation Models–
The key insight is that Orthogonal Alignment emerges naturally because it improves scaling law, and it is shown that baselines additionally incorporating cross-attention module outperform parameter-matched baselines, achieving a superior accuracy-per-model parameter.
- –
- –
- –
- Research on the Application of Image-based Gesture Recognition Technology in Classical Industries–
Through the analysis of sensor-based and vision technology, combined with practical application cases, this paper discusses its application potential and challenges in classic industries, and proposes corresponding future development directions.
- –
- –
- A Unified Knowledge-Distillation and Semi-Supervised Learning Framework to Improve Industrial Ads Delivery Systems–
A Unified framework for Knowledge-Distillation and Semi-supervised Learning for ads ranking is proposed, empowering the training of models on a significantly larger and more diverse datasets, thereby reducing overfitting and mitigating training-serving data discrepancies.
- –
- –
- –
- –
- –
- –
- –
- Quaternion Factorization Machines: A Lightweight Solution to Intricate Feature Interaction Modelling–
- A Detection Network United Local Feature Points and Components for Fine-Grained Image Classification–
- –
- Face Feature Extraction and Recognition Based on Lighted Deep Convolutional Neural Network–
The experimental results show that the feature extraction based on deep convolutional neural network in this paper can use dynamic library DLL to play an effective role in practical commercial projects.
- –
- The design and implementation of data system based on Polymorphic Graphics and text–
The key technologies of the system, such as the main function module design, three-tier architecture quick design tool, online visio file and pdf file browsing, have strong practicability and good user interface.
- An Improved Image Registration Method for Infrared and Visible Images–
In order to solve the problem of high mismatch rate of speeded up robust features (SURF) algorithm, the constraint condition of slope consistency has been used to eliminate the number of mismatch points.
- Active Learning-Based Semi-supervised Spectral Clustering Algorithm–
An active learning algorithm based on pair-wise constraints with error correction is proposed in this paper, which reduces the computational complexity of the semi-supervised algorithms based on constraints and resolves the singular problem of the pair- Wise constraints in the clustering process.
- –
- Parallelization of Eigenvalue-Based Dimensional Reductions via Homotopy Continuation–
A homotopy-based method for embedding with hundreds of thousands of data items that yields a parallel algorithm suitable for running on a distributed system by parallelizing the corresponding eigendecomposition.
- –
- –
- Research on Multi-pattern Fusion–
Experimental results indicate the method with multimodal fusion processes semantic events in sports video more effectively, as well as the unified probabilistic framework presented.
- Research on Analysis of Sports Video Multi-Pattern Fusion–
This work presents a unified probabilistic framework for sports video analysis and introduced the method based on dynamic Bayes network model, which can not only integrate multimode information for event detection and also deal with hierarchical relationship among events.
- Comparison of recent non-linear filters from graphics field–
This paper gives a brief introduction to several nonlinear filters aiming at edge-preserving smoothing and comparison about their advantages and limitations through experiments to offer an helpful starting point for researchers to select or improve them.
- Robust algorithm of multi-signals separation for wideband spectrum sensing of cognitive radios–
This paper focuses on the multi-signals separation problem while the signal number is uncertain and the noise floor spectrum is non-flat.
- Global Health Prediction Based on Vector Projection Algorithm–
A model to provide reasonable predictions on Earth’s health condition is developed and critical and tipping points in the network are found and this information can be used for the policy making.
- Low-complexity Noncoherent Iterative CPM Demodulator for FH Communication–
A novel demodulator is proposed, which reduces the complexity by apply- ing phase quantization and exploiting the phase rotational invariance property of CPM signals and provides considerable performance improvement over the existing solutions with the same com- putational complexity.
- A 3D 2-Subiteration Thinning Algorithm for Human Pose Estimation–
The experiment results show that although the intermediate results of the shape-from-silhouettes process and 3D thinning process contain much noise, the model registration method, which bases on a set of simple prior rules, can still get accurate 3D human body skeleton model.
- –
- –
- –
- Recognizing Human Action from Videos Using Histograms of Visual Words–
A new method for human action recognition from video sequences using histograms of visual words, where each frame corresponds to a "word" in a novel "bag-of-words" representation.
- The Development of the SolidWorks Standard Components Library Based on API–
A method with which you can secondary develop SolidWorks using API to make the standard components library and the SolidWorks together, which make the design more efficient when the engineer designs a product with SolidWorks.
- –
- Image Understanding with Connected Component Histograms.–
This work believes image understanding may be thought of as a process which is composed of three levels – pixel, component and con understanding, and focuses on the properties of the pixels in the image.
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.