Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks176 from public data
- Attentive Moment Retrieval in Videos351
An Attentive Cross-Modal Retrieval Network is developed that designs a memory attention mechanism to emphasize the visual features mentioned in the query and simultaneously incorporate their context, and obtains the augmented moment representation.
- Dynamic Modality Interaction Modeling for Image-Text Retrieval230
A novel modality interaction modeling network based upon the routing mechanism, which is the first unified and dynamic multimodal interaction framework towards image-text retrieval and demonstrates superiority compared with several state-of-the-art baselines.
- Context-Aware Multi-View Summarization Network for Image-Text Matching165
A novel context-aware multi-view summarization network to summarize context-enhanced visual region information from multiple views and designs an adaptive gating self-attention module to extract representations of visual regions and words.
- Online Data Organizer: Micro-Video Categorization by Structure-Guided Multimodal Dictionary Learning113
- 110
- NTIRE 2020 Challenge on Real Image Denoising: Dataset, Methods and Results106
This paper reviews the NTIRE 2020 challenge on real image denoising with focus on the newly introduced dataset, the proposed methods and their results, based on the SIDD benchmark.
- Crop Classification Based on Temporal Information Using Sentinel-1 SAR Time-Series Data105
A crop classification scheme, which made full use of multitemporal SAR backscattering responses, and achieved the best overall accuracies in both VH and VV bands is proposed.
- 100
- 86
- 76
- The Diversified Ensemble Neural Network74
Results show the proposed principled ensemble technique by constructing the so-called diversified ensemble layer to combine multiple networks as individual modules can notably improve the accuracy and stability of the original neural networks with ignorable extra time and space overhead.
- 69
- Bi-directional Heterogeneous Graph Hashing towards Efficient Outfit Recommendation54
This work devise a novel bi-directional heterogeneous graph hashing scheme, called BiHGH, towards efficient personalized outfit recommendation and adopt the bayesian personalized ranking loss for the user preference learning and design the dual similarity preserving regularization to prevent the information loss during hash learning.
- 52
- 52
- Fine-grained Textual Inversion Network for Zero-Shot Composed Image Retrieval51
A novel Fine-grained Textual Inversion Network for ZS-CIR, named FTI4CIR, which comprises two main components: fine-grained pseudo-word token mapping and tri-wise caption-based semantic regularization.
- A Survey on Video Moment Localization51
This survey paper aims to present a comprehensive review of existing video moment localization techniques, including supervised, weakly supervised, and unsupervised ones, and to discuss promising future directions for this field, in particular large-scale datasets and interpretable video moments localization models.
- Efficient Mean‐shift Clustering Using Gaussian KD‐Tree50
An efficient method that allows mean shift clustering performed on large data set containing tens of millions of points at interactive rate is proposed, using a new scheme for approximating mean shift procedure using a greatly reduced feature space.
- Bendlet Transform Based Adaptive Denoising Method for Microsection Images48
An adaptive denoising method for microsection images with Rician noise is proposed, in which the curve contour and texture can be identified as low-frequency components, which is not the case with other methods, such as the wavelet, shearlet, and so on.
- STRONG: Spatio-Temporal Reinforcement Learning for Cross-Modal Video Moment Localization48
This article innovatively contributes a spatial-temporal reinforcement learning framework that first exploits a temporal-level reinforcement learning to dynamically adjust the boundary of localized video moment instead of the traditional window segmentation strategy, which is able to accelerate the localization process.
- 48
- Multimodal Dialog System: Relational Graph-based Context-aware Question Understanding46
This paper presents a relational graph-based context-aware question understanding scheme, which enhances the user intention comprehension from local to global, and designs a sparse graph attention network to adaptively aggregate effective context information for each utterance.
- Semantic Collaborative Learning for Cross-Modal Moment Localization43
This work designs a dual-path representation network for intra-modal semantic modeling and develops a semantic collaborative network to achieve multi-granularity semantic alignment and hierarchical semantic supplement and proves that effective moment localization can be achieved based on sufficient semantic collaborative learning.
- 41
- UniSA: Unified Generative Framework for Sentiment Analysis40
This work proposes a Task-Specific Prompt method to jointly model subtasks and introduces a multimodal generative framework called UniSA, and designs novel pre-training tasks and training methods to enable the model to learn generic sentiment knowledge among subtasks to improve the model's multi-modal sentiment perception ability.
- Sparsely Grouped Multi-Task Generative Adversarial Networks for Facial Attribute Manipulation39
Experimental results show that SG-GAN can achieve comparable results with state-of-the-art methods on adequately labelled datasets while attaining a superior image translation quality on sparsely grouped datasets.
- A Survey on Video Temporal Grounding With Multimodal Large Language Model34
Current research on VTG-MLLMs is examined through a three-dimensional taxonomy: the functional roles of MLLMs, highlighting their architectural significance; training paradigms, analyzing strategies for temporal reasoning and task adaptation; and video feature processing techniques, which determine spatiotemporal representation effectiveness.
- 33
- 33
- Multigranular Visual-Semantic Embedding for Cloth-Changing Person Re-identification31
A novel multigranular visual-semantic embedding algorithm (MVSE) is proposed for cloth-changing person ReID, where visual semantic information and human attributes are embedded into the network, and the generalized features of human appearance can be well learned to effectively solve the problem of cloth- changing.
- Human voice attractiveness processing: Electrophysiological evidence30
Takeaway is that voice attractiveness processing during early stages appears to be rapid and mandatory and related to mate selection, whereas during later elaborated processing, voice attractiveness is strategic and aesthetics-based, requiring attentional resources.
- Exo2Ego: Exocentric Knowledge Guided MLLM for Egocentric Video Understanding29
This work introduces Ego-ExoClip, a pre-training dataset comprising 1.1M synchronized ego-exo clip-text pairs derived from Ego-Exo4D, together with the instruction-tuning dataset EgoIT, which is collected from multiple sources to enhance the model's instruction-following capabilities.
- A New X-ray Medical-Image-Enhancement Method Based on Multiscale Shannon–Cosine Wavelet28
An image-enhancement algorithm based on wavelet multiscale decomposition is proposed to address the shortcomings of existing single-scale image- enhancement algorithms, based on Shannon–Cosine wavelets by taking advantage of the interpolation, smoothness, tight support, and normalization properties.
- Learnable Pillar-based Re-ranking for Image-Text Retrieval25
This paper designs a neighbor-aware graph reasoning module to flexibly exploit the relations and excavate the sparse positive items within a neighborhood, and presents a structure alignment constraint to promote crossmodal collaboration and align the asymmetric modalities.
- Efficient Brain Tumor Segmentation with Lightweight Separable Spatial Convolutional Network24
The Separable Spatial Convolutional Network is introduced, an innovative model that refines the U-Net architecture to achieve efficient brain tumor segmentation with minimal computational cost and incorporates a separable spatial attention mechanism.
- Hawkes-Enhanced Spatial-Temporal Hypergraph Contrastive Learning Based on Criminal Correlations24
Hawkes-enhanced Spatial-Temporal Hypergraph Contrastive Learning framework (HCL), which mines the aforementioned correlations via two specific strategies via two specific strategies, and shows promising capacities from four aspects, i.e., superiority, transferability, effectiveness, and sensitivity.
- 24
- 24
- Market2Dish24
A word-class interaction based text classification model is employed to learn the fine-grained similarity between sparse health features on the social media platforms and pre-defined health concepts, and then a category-aware hierarchical memory network based recommender is introduced to learning the user-recipe interactions for better food recommendations.
- 24
- Rapid Nanopore Assay for Carbapenem-Resistant Klebsiella pneumoniae23
It is proved that only 4 h of CRKP culture time was needed for nanopore assay to distinguish theCRKP and CSKP, and this new method has the potential application in the fast screening of drug resistance in clinical microorganism samples.
- 22
- 22
- Convolutional Recurrent Neural Networks with a Self-Attention Mechanism for Personnel Performance Prediction22
A hybrid convolutional recurrent neural network model based on self-attention mechanism is presented, which can automatically learn discriminative features and capture global contextual information from personnel performance data and is applied to a real case of personnel performance prediction.
- Object-Shot Enhanced Grounding Network for Egocentric Video21
OSGNet, an Object-Shot enhanced Grounding Network for egocentric video, extracts object information from videos to enrich video representation, particularly for objects highlighted in the textual query but not directly captured in the video features.
- Privacy protection for 3D point cloud classification based on an optical chaotic encryption scheme21
Various security analysis results validate that the proposed privacy protection scheme has high security level and good privacy protection effect for 3D point cloud classification.
- Adversarial Separation Network for Speaker Recognition21
This study proposes the adversarial separation network (AS-Net) to protect the speaker recognition system against adversarial attacks, featured by its ability to separate adversarial perturbation from the test speech to restore the natural clean speech.
- DDIFN: A Dual-discriminator Multi-modal Medical Image Fusion Network20
This work presents a novel end-to-end unsupervised network to fuse multi-modal medical images, composed of a generator and two symmetrical discriminators, and defines the content loss function, which preserves the gradient information and pixel activity of source images.
- 20
- A deep learning model for drug screening and evaluation in bladder cancer organoids20
A novel ACU2Net model, AU2Net (Attention and Cross U2Net), which can provide more accurate segmentation results from organoid images and can improve the efficiency of drug screening evaluation using organoids.
- Joint Feature Enhancement and Speaker Recognition with Multi-Objective Task-Oriented Network19
An asynchronous subregion optimization approach for the joint training of feature enhancement and speaker embedding neural networks and channel-wise feature concatenation to address the distortion of speaker information that is caused by enhancement loss is proposed.
- Novel Mobilizable Genomic Island GEI-D18A Mediates Conjugational Transfer of Antibiotic Resistance Genes in the Multidrug-Resistant Strain Rheinheimera sp. D1819
This study represents the first characterization of a mobilizable antibiotic resistance island in a species of Rheinheimera and provides evidence that RheInheimera spp.
- Unlocking the Power of Multimodal Learning for Emotion Recognition in Conversation18
This work proposes fine-grained adaptive gradient modulation, a plug-in approach to rebalance the gradients of modalities, and shows that it improves the performance of all baseline models and outperforms existing plug- in methods.
- 18
- Focal and Composed Vision-semantic Modeling for Visual Question Answering18
This paper proposes a focal and composed vision-semantic modeling method, which is a trainable end-to-end model for better vision- Semantic redundancy removal and compositionality modeling, and introduces the LENA cell, a plug-and-play reasoning module, which removes redundant semantic by a focal mechanism in the first step.
- Adapting Generative Pretrained Language Model for Open-domain Multimodal Sentence Summarization17
A prompt-guided image encoding module to extract the source image feature and an explicit source critical token learning module to directly capture the critical tokens of the source sentence with the reference of thesource image, where the predicted tokens can be easily understood by humans and hence improve the interpretability of the model.
- Cross-Modal Audio-Visual Co-Learning for Text-Independent Speaker Verification17
This paper proposes a cross-modal speech co-learning paradigm based on an audio-visual pseudo-siamese structure to learn the modality-transformed correlation and achieves around 60% and 20% average relative performance improvement over baseline unimodal and fusion systems, respectively.
- Advancing Video Question Answering with a Multi-modal and Multi-layer Question Enhancement Network16
The multi-modal and multi-layer question enhancement network is introduced, a groundbreaking framework emphasizing nuanced question understanding and underscoring the importance of advanced question comprehension in VideoQA.
- Replay attack detection using variable-frequency resolution phase and magnitude features16
A novel feature extraction method that leverages both the phase-based and magnitude-based features and fully utilizes the subband information and the complementary information from the phase and magnitude spectra is proposed.
- 15
- 15
- Sentiment-Enhanced Graph-Based Sarcasm Explanation in Dialogue14
A lexicon-guided utterance sentiment inference module, where a heuristic utterance sentiment refinement strategy is devised, and a context-sentiment graph to comprehensively model the semantic relations among the utterances, utterance sentiments, and video-audio sentiments, to facilitate sarcasm explanation generation are proposed.
- 13
- 13
- Photothermal effects of CuS-BSA nanoparticles on H22 hepatoma-bearing mice13
Results indicated that the CuS-BSA NPs had good dispersibility in three different solvents and had a strong absorption peak at 980 nm, and are promising photothermal agents in the photothermal therapy of cancer.
- 12
- 12
- 12
- 12
- FiRE: Enhancing MLLMs with Fine-Grained Context Learning for Complex Image Retrieval11
An automated fine-grained multimodal quintuple dataset construction pipeline and a novel two-stage fine-grained multimodal fine-tuning strategy to sequentially enhance the model's context understanding and query-target alignment capabilities, thereby improving retrieval performance.
- 11
- 11
- MRISNet:Deep-learning-based Martian instance segmentation against blur11
This work tries to solve the problems of rock sampling of rovers by developing a deep learning network, dubbed the Mars Rover Instance Segmentation Network (MRISNet), and uses generative adversarial networks to mitigate image degradation caused by blur and improve the clarity of image texture.
- 11
- Question Tagging via Graph-guided Ranking10
A graph-guided topic ranking model to tag questions in the cQA sites appropriately is devised and a multi-level context-modeling-based question encoder is designed to obtain the enhanced question representation.
- 10
- Observation of the middle intestinal tight junction structure, cloning and studying tissue distribution of the four Claudin genes of the grass carp (Ctenopharyngodon idellus)10
This study revealed the existence of the tight junction (TJ) in the middle intestinal and obtained the genetic information of Claudin-3, Claud in-15a, Claudinb and Claudinc of grass carp, aiming to found the molecular biology basis for the further study of the intestinal barrier function of Grass carp.
- 9
- 9
- Significance of relative phase features for shouted and normal speech classification8
The experimental findings show that the RP, LPAES-RP, and LPR-RP features provide promising results for the detection of shouted speech, and the proposed GRP feature can provide better results than those of the standard mel-frequency cepstral coefficient (MFCC) feature.
- Uncovering Hidden Connections: Iterative Search and Reasoning for Video-grounded Dialog8
This work presents an iterative tracking and reasoning strategy that amalgamates a textual encoder, a visualencoder, and a generator that is adept at gleaning nuances from dialog history that are pivotal to deciphering the posed questions.
- 8
- Detecting Spammer Communities Using Network Structural Features8
This work portrays reviewers as a comment-based reviewer network through a new collusion similarity measure, divides reviewers into different communities using an effective community detection method and separate spammer communities from normal reviewer communities through network structure.
- 7
- Deep Spectro-temporal Artifacts for Detecting Synthesized Speech7
This paper provides an overall assessment of track 1 (Low-quality Fake Audio Detection) and track 2 (Partially Fake Audio detection) and explained the effectiveness of the proposed greedy fusion strategy.
- Learning the Incremental Warp for 3D Vehicle Tracking in LiDAR Point Clouds7
This study leveraged a powerful target discriminator and an accurate state estimator to robustly track target objects in challenging point cloud scenarios and proposed a state estimation subnetwork that aims to learn the incremental warp for updating the coarse target state.
- 7
- Fast SVM Trained by Divide-and-Conquer Anchors7
This paper proposes to choose the representative points which are noted as anchors obtained from non-negative matrix factorization in a divide-and-conquer framework, and use the anchors to train an approximate SVM, and shows that the solving the DCA-SVM can yield an approximate solution close to the primal SVM.
- 6
- Why Representation Engineering Works: A Theoretical and Empirical Study in Vision-Language Models6
This work makes the first attempt to extend RepE to VLMs, analyzing how multimodal representations are preserved and transformed and develops a theoretical framework that explains the stability of neural activity across layers using the principal eigenvector, uncovering the underlying mechanism of RepE.
- Convolutional graph neural networks-based research on estimating heavy metal concentrations in a soil-rice system6
The usefulness of combining graph learning and convolutional calculations in the prediction of heavy metal concentrations is demonstrated and provides a new perspective to build multidimensional and multi-scale complex ecosystem models.
- 6
- 6
- Robust Object Tracking with Occlusion Handling based on Local Sparse Representation6
This paper proposes a novel tracking method taking advantage of local sparse representation to detect occlusions during the tracking sequence and demonstrates that the proposed tracking algorithm performs favorably against several state-of-the-art methods.
- Video Timeline Modeling For News Story Understanding5
This paper curates a realistic benchmark dataset, YouTube-News-Timeline, consisting of over $12$k timelines and $300$k YouTube news videos, and proposes a set of quantitative metrics to comprehensively evaluate and compare methodologies.
- 5
- Enhancing model accuracy and calibration efficiency with image based pattern selection using machine learning techniques5
This new pattern selection solution, connected with ASML Brion’s MXP (Metrology of eXtreme Performance) contour fidelity gauges and superior, accurate Newron (deep learning) resist model, fulfills the advanced technology node demands for OPC modeling, thus offering full chip prediction power.
- Information Structures in an Incomplete Interval-Valued Information System5
Information structures in an IIVIS are obtained and the dependence and information distance are presented, and group and lattice characterizations of information structures in an IIVIS are studied.
- 5
- 4
- A survey of micro-video analysis4
A review of the studies on micro-video analysis in terms of micro-video classification, prediction, and recommendation, and an overview of the methods, features, datasets, and evaluation metrics relating to these studies.
- 4
- 4
- 4
- 4
- Logistic Regression Parameter Estimation Based on Parallel Matrix Computation4
By applying a new model based on parallel matrix computing methods, the bottleneck of computing for logistic regression algorithm was overcome successfully and Experimental results proved that the new computing model can achieve nearly linear speedup.
- Universal Relocalizer for Weakly Supervised Referring Expression Grounding3
The Universal Relocalizer is introduced, a novel approach designed for weakly supervised referring expression grounding that strives to pinpoint a target proposal that corresponds to a specific query, eliminating the need for region-level annotations during training.
- Vision-guided and Mask-enhanced Adaptive Denoising for Prompt-based Image Editing3
A Vision-guided and Mask-enhanced Adaptive Editing (ViMAEdit) method with three key novel designs to leverage image embeddings as explicit guidance to enhance the conventional textual prompt-based denoising process, where a CLIP-based target image embedding estimation strategy is introduced.
- Empowering GNNs via Edge-Aware Weisfeiler-Leman Algorithm3
This work proposes a general and provably powerful GNN framework that preserves the scalability of message passing scheme, and proposes the NC-GNN framework as a differentiable neural version of NC-1-WL.
- 3
- 3
- 3
- Uncovering Hidden Connections: Iterative Search and Reasoning for Video-grounded Dialog2
An iterative search and reasoning framework composed of a textual encoder, a visual encoder, and a generator that adopts a path search and aggregation strategy to identify key cues in the dialog history that are essential for understanding the current question.
- Liquid Level Detection of Polytetrafluoroethylene Emulsion Rotary Vibrating Screen Device Based on TransResNet2
This research presents a deep-learning-oriented video surveillance model for the intelligent level detection of vibrating screens, waste drums, and emulsion outlets, effectively addressing the limitations of traditional methods.
- 2
- 2
- Harnessing Representative Spatial-Temporal Information for Video Question Answering2
A spatial-temporal representative visual exploitation network for video question answering is proposed, which enhances the understanding of the video by merely summarizing representative visual information, and advance adaptive attention based on uncertainty estimation.
- SVEN_5003 is a Major Developmental Regulator in Streptomyces venezuelae2
This study identified sven_5003 in Streptomyces venezuelae as a major new developmental regulatory gene, the deletion of which leads to a bald phenotype, typical of bld mutants, under multiple growth conditions.
- A Baseline for ViCo Conversational Head Generation Challenge2
A three-stage method to generate high-quality and lip-sync talking head videos with natural head poses wins 3rd place and People's Selection Award in the ViCo conversational head generation challenge.
- 2
- 2
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- PatchRLNet: A Framework Combining a Vision Transformer and Reinforcement Learning for The Separation of a PTFE Emulsion and Paraffin1
An automated detection framework named PatchRLNet is proposed, which leverages a combination of a vision transformer and reinforcement learning, and represents the first instance of automated detection applied to the separation of the PTFE emulsion and paraffin.
- 1
- Complex Shaped Detection and Reconstruction Algorithm Considering Uncertain Factors1
This paper designs a complex shaped detection and reconstruction algorithm based on linear structured light that performs the visually assisted localization of the ROI area of the welding groove in a complex environment, and completes the extraction of the characteristic area ofThe welding groove and the interpolation of missing parts based on the slope method.
- 1
- Towards Realistic Conversational Head Generation: A Comprehensive Framework for Lifelike Video Synthesis1
A comprehensive method that combines audio-driven 3DMM parameter prediction with a rendering network to generate high-fidelity lip-sync talking head videos that achieved first place in the talking head generation track of the challenge and was honored with the People's Selection Award.
- 1
- 1
- Toward community answer selection by jointly static and dynamic user expertise modeling1
This paper proposes an inductive strategy to aggregate the social influence of neighbors and introduces the explicit topic interest of users and captures the context-based personal interest by weighing the activation of each topic.
- HCPC: A New Parsimonious Clustering Method based on Hierarchical Characters for Morphological Phylogenetic Reconstruction1
This study proposes a phylogenetic clustering method based on hierarchical characters that is able to infer the shared ancestral relationships among species and uses a simulated annealing algorithm to heuristically search the phylogenetic tree based on the parsimony criterion.
- 1
- An Unsupervised Approach for 3D Face Reconstruction from a Single Depth Image1
A convolutional encoder network to learn a mapping function from a noisy depth image to a 3D expressive facial model is proposed and trained in an unsupervised manner by exploiting the consistent fitting of the 3D mesh and the depth image.
- 1
- A clinical study of neurological signs of medullary infarction evaluated with expanded-National Institutes of Health Stroke Scale1
The e-NIHSS could improve the sensitivity of NIHSS in evaluating the neurological signs of medullary infarction, which is better in evaluating LMI than in evaluating MMI.
- –
- –
- –
- Egocentric Action Recognition with Retrieval-Augmented Learning–
REAR is proposed, a Retrieval-augmented framework for EAR that leverages external third-person (exocentric) videos as auxiliary knowledge—without requiring synchronized ego-exo pairs.
- –
- –
- –
- A Novel Multi-View Perception and Shrinkage Aggregation Network for Inharmonious Region Localization–
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Editorial–
This special issue seeks recent advances in the deep learning-based multimedia analytics and relatively new areas and proposes a hybrid features and semantic reinforcement network to detect and localize image manipulation regions at the pixel level, including copy-move, splicing, and object removal.
- –
- –
- –
- –
- –
- –
- –
- –
- An Instant Interest Oriented Recommendation Agent Based on SVDD and Schedule–
In order to detect the instant interest in mobile commerce environments, an interactive recommender agent was proposed based on schedules and the algorithm of support vector data description and the result proved that the agent was effective.
- Analysis on the Structure and the Hidden Markov Model–
The second frame differential method to detect the shot of soccer video is constructed and the test results show that the precision of this method is higher in close-up and misjudgment slow motion, however, it has certain leak detection in slow motion replay of high speed movement.
- Classification by Essential Emerging Patterns–
This work proposes a novel classification method, CEEP (Classification by Essential Emerging Patterns), which uses a special kind of EPs, called essential Emerging Patterns (eEPs), and a growth-rate-based scoring method to construct classifiers.
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 2026-10-12. 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.