Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- AN Jin-wenSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works284 from public data
- 2,985
- Spatio-Temporal LSTM with Trust Gates for 3D Human Action Recognition1,414
This paper introduces new gating mechanism within LSTM to learn the reliability of the sequential input data and accordingly adjust its effect on updating the long-term context information stored in the memory cell, and proposes a more powerful tree-structure based traversal method.
- 1,385
- DVC: An End-To-End Deep Video Compression Framework895
This paper proposes the first end-to-end video compression deep model that jointly optimizes all the components for video compression, and shows that the proposed approach can outperform the widely used video coding standard H.264 in terms of PSNR and be even on par with the latest standard MS-SSIM.
- 760
- 632
- A Siamese Long Short-Term Memory Architecture for Human Re-identification596
A novel siamese Long Short-Term Memory (LSTM) architecture that can process image regions sequentially and enhance the discriminative capability of local feature representation by leveraging contextual information.
- 553
- Skeleton-Based Action Recognition Using Spatio-Temporal LSTM Network with Trust Gates540
A new gating mechanism within LSTM module is introduced, with which the network can learn the reliability of the sequential data and accordingly adjust the effect of the input data on the updating procedure of the long-term context representation stored in the unit's memory cell.
- Salient Object Detection via Integrity Learning480
This work designs a novel integrity learning for SOD, which outperforms the baseline methods in terms of a wide range of metrics and introduces an integrity channel enhancement (ICE) component with the goal of enhancing feature channels that highlight the integral salient objects, while suppressing the other distracting ones.
- 466
- 438
- 414
- 407
- 406
- 380
- 360
- 356
- FVC: A New Framework towards Deep Video Compression in Feature Space354
This work proposes a feature-space video coding network (FVC) by performing all major operations (i.e., motion estimation, motion compression, motion compensation and residual compression) in the feature space by using the auto-encoder style network.
- Domain adaptation from multiple sources via auxiliary classifiers339
A new data-dependent regularizer based on smoothness assumption into Least-Squares SVM (LS-SVM), which enforces that the target classifier shares similar decision values with the auxiliary classifiers from relevant source domains on the unlabeled patterns of the target domain.
- 334
- 306
- 294
- 261
- 254
- 243
- 220
- 216
- 207
- 204
- VoxelContext-Net: An Octree based Framework for Point Cloud Compression196
To alleviate the distortion from the octree construction procedure, a voxel context based 3D coordinate refinement method is proposed to produce more accurate reconstructed point cloud at the decoder side, which is applicable to both static and dynamic point cloud compression.
- Ranking with local regression and global alignment for cross media retrieval191
A novel ranking algorithm, namely ranking with Local Regression and Global Alignment (LRGA), which learns a robust Laplacian matrix for data ranking and a unified objective function to globally align the local models from all the data points so that an optimal ranking value can be assigned to each data point.
- 178
- 178
- Content Adaptive and Error Propagation Aware Deep Video Compression176
This work proposes a content adaptive and error propagation aware video compression system that outperforms the state-of-the-art learning based video codecs on benchmark datasets without increasing the model size or decreasing the decoding speed.
- Human-Centric Spatio-Temporal Video Grounding With Visual Transformers163
This work introduces a novel task – Human-centric Spatio-Temporal Video Grounding (HC-STVG), which aims to localize a spatio-temporal tube of the target person from an untrimmed video based on a given textural description.
- 162
- 162
- Improving Deep Video Compression by Resolution-Adaptive Flow Coding149
A new framework called Resolution-adaptive Flow Coding (RaFC) is proposed to effectively compress the flow maps globally and locally, in which multi-resolution representations are used for both the input flow maps and the output motion features of the MV encoder.
- 147
- VDM-DA: Virtual Domain Modeling for Source Data-Free Domain Adaptation146
This work deals with the source data-free unsupervised domain adaptation problem and proposes a novel approach referred to as Virtual Domain Modeling for Domain Adaptation (VDM-DA), in which the virtual domain acts as a bridge between the source and target domains.
- 144
- 144
- 143
- 140
- 136
- 136
- Dependency Exploitation: A Unified CNN-RNN Approach for Visual Emotion Recognition135
This paper proposes a unified CNNRNN model to predict the emotion based on the fused features from different levels by exploiting the dependency among them and leverages convolutional neural network with multiple layers to extract different levels of features within a multi-task learning framework.
- Back-tracing Representative Points for Voting-based 3D Object Detection in Point Clouds131
This work introduces a new 3D object detection method, named as Back-tracing Representative Points Network (BRNet), which generatively back-traces the representative points from the vote centers and also revisits complementary seed points around these generated points, so as to better capture the fine local structural features surrounding the potential objects from the raw point clouds.
- Coarse-To-Fine Deep Video Coding with Hyperprior-Guided Mode Prediction129
A coarse-to-fine (C2F) deep video compression framework equipped with the new hyperprior-guided mode prediction methods achieves the state-of-the-art performance on HEVC, UVG and MCL-JCV datasets.
- 121
- 120
- Recent Advances in Transfer Learning for Cross-Dataset Visual Recognition119
This survey categorises the cross-dataset recognition into 17 problems based on a set of carefully chosen data and label attributes to provide a systematic approach and a reference for a machine-learning practitioner to categorise a real problem and to look up for a possible solution accordingly.
- Diffusion Model is Secretly a Training-Free Open Vocabulary Semantic Segmenter115
The potential of generative text-to-image diffusion models as highly efficient open-vocabulary semantic segmenters are uncovered, and a novel training-free approach named DiffSegmenter is introduced that achieves impressive results for open-vocabulary semantic segmentation.
- 109
- Deep Kalman Filtering Network for Video Compression Artifact Reduction108
This paper model the video artifact reduction task as aKalman filtering procedure and restore decoded frames through a deep Kalman filtering network and builds a recursive filtering scheme based on the Kalman model.
- 107
- Semisupervised Dimensionality Reduction and Classification Through Virtual Label Regression107
The proposed approach propagates the label information from labeled to unlabeled data with a well-designed mechanism of random walks, in which outliers are effectively detected and the obtained virtual labels of unlabeling data can be well encoded in a weighted regression model.
- 106
- Influence Spreading Path and Its Application to the Time Constrained Social Influence Maximization Problem and Beyond100
It is shown that the problem is NP-hard, and the monotonicity and submodularity of the time constrained influence spread function is proved, and a greedy algorithm is developed based on this, which is developed to improve the algorithm scalability.
- 98
- 96
- 96
- Detection of Sudden Pedestrian Crossings for Driving Assistance Systems93
This paper proposes a three-level coarse-to-fine video-based framework that detects partially visible pedestrians just as they enter the camera view, with low false alarm rate and high speed and yields a performance better than that of state-of-the-art pedestrian detection while running at a frame rate of 55 fps.
- 89
- 89
- 88
- 87
- 84
- Dividing and Aggregating Network for Multi-view Action Recognition83
A new Dividing and Aggregating Network (DA-Net) for multi-view action recognition based on the conditional random field (CRF) formulation to pass message among view-specific representations from different branches to help each other.
- Weighted Block-Sparse Low Rank Representation for Face Clustering in Videos83
A new method called weighted block-sparse low rank representation (WBSLRR) is proposed which considers the available prior knowledge while learning a low rank data representation, and also develops a simple but effective approach to obtain the clustering result of faces.
- Side-Information based Linear Discriminant Analysis for Face Recognition80
Side-Information based Linear Discriminant Analysis (SILD), in which the within-class and between-class scatter matrices are directly calculated by using the side-information, is proposed, and it is theoretically prove that the SILD method is equivalent to FLDA when the full class label information is available.
- Model Compression Using Progressive Channel Pruning79
This work proposes a simple but effective channel pruning framework called Progressive Channel Pruning (PCP) to accelerate Convolutional Neural Networks (CNNs) and extends the PCP framework to prune channels for the deep transfer learning methods like Domain Adversarial Neural Network (DANN), in which the data distribution mismatch is effectively reduced.
- 79
- Flowing on Riemannian Manifold: Domain Adaptation by Shifting Covariance76
A new unsupervised domain adaptation method called domain adaptation by shifting covariance (DASC) by utilizing the principal components of each covariance matrix for object recognition without requiring any labeled samples from the target domain is proposed.
- 76
- Sliding Window and Regression Based Cup Detection in Digital Fundus Images for Glaucoma Diagnosis75
A machine learning framework based on sliding windows for glaucoma diagnosis based on an epsilon-SVR (support vector regression) model based on non-linear radial basis function (RBF) kernels and final decisions are made with a non-maximal suppression (NMS) method.
- 72
- 71
- 70
- Automatic Grading of Nuclear Cataracts from Slit-Lamp Lens Images Using Group Sparsity Regression70
A new feature for cataract grading is introduced together with a group sparsity-based constraint for linear regression, which performs feature selection, parameter selection and regression model training simultaneously and has potential for better clinical management of the disease.
- 69
- 67
- Adaptive discriminant learning for face recognition67
Adaptive Discriminant Analysis (ADA) is proposed in which the within-class scatter matrix of each enrolled subject is inferred using his/her single sample, by leveraging a generic set with multiple samples per person.
- 66
- 64
- Exploiting Privileged Information from Web Data for Image Categorization64
A new image categorization method is proposed by incorporating the textual features extracted from the surrounding textual descriptions (tags, captions, categories, etc.) as privileged information and simultaneously coping with noise in the loose labels of training web images.
- Efficient Reconstruction-Based Optic Cup Localization for Glaucoma Screening64
This work presents a reconstruction-based learning technique to localize the optic cup in fundus images for glaucoma screening that can be formulated as a closed-form solution without any search, which leads to highly efficient and 100% repeatable computation.
- 64
- 64
- Initialization Independent Clustering With Actively Self-Training Method62
An actively self-training clustering method, in which the samples are actively selected as training set to minimize an estimated Bayes error, and then explore semisupervised learning to perform clustering.
- 61
- Channel Pruning Guided by Classification Loss and Feature Importance59
This work proposes a new layer-by-layer channel pruning method called Channel Pruning guided by classification Loss and feature Importance (CPLI), which additionally take the classification loss into account in the channel pruned process.
- 59
- 58
- Dimensionality-Dependent Generalization Bounds for k-Dimensional Coding Schemes58
A dimensionality-dependent generalization bound is derived for k-dimensional coding schemes by bounding the covering number of the loss function class induced by the reconstruction error, and it is shown that this bound can be tighter than previous results because it avoids inducing the worst-case upper bound on k of theloss function.
- 58
- 57
- Dense Video Captioning Using Graph-Based Sentence Summarization57
A graph-based partition-and-summarization (GPaS) framework for dense video captioning within two stages that effectively exploits the relationship between semantic words for summarization and achieves this goal by treating semantic words as the nodes in a graph.
- 57
- Beyond Spatial Pyramids: A New Feature Extraction Framework with Dense Spatial Sampling for Image Classification57
This work proposes a new learning algorithm, called Generalized Adaptive lp-norm Multiple Kernel Learning (GA-MKL), to learn an adapted robust classifier based on multiple base kernels constructed from image features and multiple sets of pre-learned classifiers of all the classes.
- 56
- Self-Paced Collaborative and Adversarial Network for Unsupervised Domain Adaptation54
This paper proposes a new unsupervised domain adaptation approach called Collaborative and Adversarial Network (CAN), which uses the domain-collaborative and domain-adversarial learning strategies for training the neural network.
- 54
- 54
- 52
- 52
- 52
- 51
- 51
- 51
- 50
- Inception Convolution with Efficient Dilation Search49
A new type of dilated convolution is proposed, where the convolution operations have independent dilation patterns among different axes, channels and layers, and this method achieves consistent performance gains for image recognition, object detection, instance segmentation, human detection, and human pose estimation.
- 49
- 49
- 48
- Large expert-curated database for benchmarking document similarity detection in biomedical literature search47
The RElevant LIterature SearcH consortium consisting of more than 1500 scientists from 84 countries, who have collectively annotated the relevance of over 180 000 PubMed-listed articles with regard to their respective seed (input) article/s, shows that the three representative baseline methods used to generate recommended articles for evaluation had similar overall performances.
- 46
- 45
- Efficient Optic Cup Detection from Intra-image Learning with Retinal Structure Priors45
A superpixel based learning framework based on retinal structure priors for glaucoma diagnosis that proposes processing of the fundus images at the superpixel level, which leads to features more descriptive and effective than those employed by pixel-based techniques, while yielding significant computational savings over methods based on sliding windows.
- 44
- Joint Optimization Toward Effective and Efficient Image Search44
This paper systematically investigates the image search performance variation with respect to a few factors of the bag-of-words model, and proposes an effective and efficient image search scheme, which is jointly optimized in terms of effectiveness and efficiency.
- Show, Tell and Summarize: Dense Video Captioning Using Visual Cue Aided Sentence Summarization43
This work formulate the dense video captioning task as a visual cue aided sentence summarization problem and proposes a new two stage Long Short Term Memory (LSTM) approach equipped with a new hierarchical attention mechanism to summarize all generated sentences as one descriptive sentence with the aid of visual features.
- 41
- 40
- Retrieval based interactive cartoon synthesis via unsupervised bi-distance metric learning40
A gesture recognition method for cartoon character images with two applications, namely content-based cartoon image retrieval and cartoon clip synthesis, that learns the optimal distance metric from the heterogeneous distance metrics derived from intra-features and inter-features.
- 40
- 39
- Cross-media retrieval using query dependent search methods39
This work constructs an MMD semi-semantic graph (MMDSSG) by jointly analyzing the heterogeneous multimedia data and proposes different learning methods of relevance feedback (RF) to improve the performance of cross-media retrieval.
- ICMH-Net: Neural Image Compression Towards both Machine Vision and Human Vision38
This work introduces a novel approach to Partition, Transmit, Reconstruct, and Aggregate the latent representation of images to balance the optimizations for both human vision quality and machine vision tasks simultaneously, and creates a latent representation predictor that dynamically manages the bit-rate cost for machine Vision tasks.
- Exploiting Privileged Information from Web Data for Action and Event Recognition36
This work proposes new learning frameworks to train robust classifiers for action and event recognition by using freely available web videos as training data, and proposes a new framework called multi-instance learning with privileged information (MIL-PI), together with three new MIL methods that can additionally reduce the data distribution mismatch between training and test videos.
- Image Classification With Densely Sampled Image Windows and Generalized Adaptive Multiple Kernel Learning36
A framework for image classification that extends beyond the window sampling of fixed spatial pyramids and is supported by a new learning algorithm and introduces generalized adaptive ℓp-norm multiple kernel learning (GA-MKL) to learn a robust classifier based on multiple base kernels constructed from the new image features and multiple sets of prelearned classifiers from other classes.
- 35
- 35
- 34
- Neural Video Compression with Spatio-Temporal Cross-Covariance Transformers33
This work proposes a new 3D-based transformer module: Spatio-Temporal Cross-Covariance Transformer (ST-XCT), which is integrated into various key coding components of NVC, such as feature extraction, frame reconstruction, and entropy modeling, demonstrating its generalizability.
- 33
- 33
- DTMiner: identification of potential disease targets through biomedical literature mining33
A reliable and efficient framework that takes large biomedical literature repositories as inputs, identifies credible relationships between diseases and genes, and presents possible genes related to a given disease and possible diseases related toA given gene is proposed.
- 33
- Finding Correspondence from Multiple Images via Sparse and Low-Rank Decomposition33
This work proposes a robust solution to the problem of finding the correspondence from multiple images by exploiting the priors that the rank of the ordered patterns from a set of linearly correlated images should be lower than that of the disordered patterns, and the errors among the reordered patterns are sparse.
- 32
- Progressive Cross-Stream Cooperation in Spatial and Temporal Domain for Action Localization32
This work proposes a new progressive cross-stream cooperation (PCSC) framework that improves all three levels of tasks of spatio-temporal action localization, and proposes a new message passing approach to pass information from one stream to another stream, which also leads to better action detection and segment detection models.
- Automatic glaucoma diagnosis through medical imaging informatics32
AGLAIA-MII demonstrates for the first time the capability of integrating patients' personal data, medical retinal image and genome information for automatic glaucoma diagnosis and screening in a large dataset from a population study and paves the way for a holistic approach for automatic objective glaucoma diagnosis and screening.
- 32
- Using large-scale web data to facilitate textual query based retrieval of consumer photos32
A (quasi) real-time textual query based personal photo retrieval system by leveraging millions of web images and their associated rich textual descriptions (captions, categories, etc.) to improve the photo retrieval performance.
- Bridging Synthetic and Real Images: A Transferable and Multiple Consistency Aided Fundus Image Enhancement Framework31
This work proposes an end-to-end optimized teacher-student framework to simultaneously conduct image enhancement and domain adaptation, and proposes a novel multi-stage multi-attention guided enhancement network (MAGE-Net) as the backbones of the teacher and student network.
- 31
- 31
- Data-Driven Information Extraction from Chinese Electronic Medical Records31
A data-driven framework that takes unstructured free text narratives in Chinese Electronic Medical Records as input and converts them into structured time-event-description triples, where the description is either an elaboration or an outcome of the medical event.
- 30
- 29
- Slow Motion Matters: A Slow Motion Enhanced Network for Weakly Supervised Temporal Action Localization28
A novel framework termed Slow Motion Enhanced Network (SMEN) is proposed to improve the ability of a WTAL network by compensating its sensitivity on slow-motion action segments by improving the ability to localize actions in untrimmed videos with only weak supervision information.
- Progressive Modality Cooperation for Multi-Modality Domain Adaptation28
This work proposes a new generic multi-modality domain adaptation framework called Progressive Modality Cooperation (PMC) to transfer the knowledge learned from the source domain to the target domain by exploiting multiple modality clues under the MMDA and MMDA-PI settings.
- 28
- 28
- Where your photo is taken: Geolocation prediction for social images27
A unified framework to suggest geolocations for images is proposed, which combines the information from both image tags and the user profile, and is able to achieve very high accuracy for images from users who have done geotagging.
- Operational transformation for dependency conflict resolution in real-time collaborative 3D design systems27
This work is the first to extend OT capability to dependency conflict resolution, and OT application scope from 1D/2D to 3D applications, thus advancing the state-of-the-art of OT in both theory and practical application.
- 27
- 27
- 25
- 25
- 25
- Group-aware Parameter-efficient Updating for Content-Adaptive Neural Video Compression24
This work introduces a parameter-efficient delta-tuning strategy, which is achieved by integrating several light-weight adapters into each encoding component by using both serial and parallel configuration, and introduces a group-aware approach for updating encoder parameters.
- 24
- Patch Distribution Compatible Semisupervised Dimension Reduction for Face and Human Gait Recognition24
- Introduction to special issue on social media24
This special issue is initiated to provide a platform for presenting results from leading edge research in understanding and exploring social media, as well as presenting the next-generation technologies for multimedia understanding.
- 24
- Cross-Domain Detection Transformer Based on Spatial-Aware and Semantic-Aware Token Alignment23
This paper takes advantage of the characteristics of cross-attention as used in the detection transformer and proposes spatial-aware token alignment (SpaTA) and semantic-aware token alignment (SemTA) strategies to guide the token alignment across domains.
- 23
- 23
- 23
- Coupled Subspaces Analysis23
CSA effectively alleviates the curse of dimensionality dilemma and small sample size problem existed in face recognition problem and explains the superiority of 2DPCA compared with PCA.
- 21
- 20
- 20
- 20
- 20
- 20
- Guest Editor’s Introduction to the Special Issue on Domain Adaptation for Vision Applications19
This work focuses on domain adaptation in the area of multimedia collections, where users may be interested in retrieving and organizing their own multimedia collections of images and videos at the semantic level, but may be reluctant to put forth the effort to annotate their photos and videos by themselves.
- Self‐supervised learning for accelerated 3D high‐resolution ultrasound imaging18
A new deep learning-based algorithm for reconstructing high-resolution HR 3D US images from sparely acquired 2D images only reliant on the acquired sparsely distributed 2D images is proposed.
- An Improved TFIDF Algorithm in Text Classification18
From the subsequent evaluations using 10-fold cross-validation, it can see the CTF-IDF greatly improves the accuracy of text classification.
- 18
- Social Media Modeling and Computing17
This timely text/reference presents the latest advances in various aspects of social media modeling and social media computing research with contributions from an international selection of preeminent experts in the field.
- 17
- 16
- Improving Weakly Supervised Temporal Action Localization by Exploiting Multi-Resolution Information in Temporal Domain16
This work proposes a two-stage approach to generate high-quality frame-level pseudo labels by fully exploiting multi-resolution information in the temporal domain and complementary information between the appearance and motion streams, and uses a Progressive Temporal Label Refinement framework to iteratively refine the pseudo labels.
- Attribute based salient image extrema detection algorithm16
A new salient image extrema detection algorithm applying attribute based morphological analysis is presented and experimental results show the effectiveness and applicability of the proposed algorithm.
- 15
- Global processing provides malignancy evidence complementary to the information captured by humans or machines following detailed mammogram inspection15
The gist signal provided malignancy evidence with no or weak associations with the information captured by humans in normal radiologic reporting, which involves detailed mammogram inspection.
- Domain Adaptive Fisher Vector for Visual Recognition15
A Domain Adaptive Fisher Vector (DAFV) method is proposed, which learns a transformation matrix to select the domain invariant components of Fisher vectors and simultaneously solves a regression problem for visual recognition tasks based on the transformed features.
- 15
- 14
- 14
- 14
- 13
- 13
- 13
- 13
- 12
- 12
- Self-calibration of multi-camera networks without feature correspondence between different cameras11
An automatic calibration approach for multi-camera networks is proposed to calibrate the intrinsic parameters of each camera and the extrinsic parameters between different cameras by a self-calibration method from the motion of feature points.
- 10
- 10
- 10
- 3D Gaussian Splatting Data Compression with Mixture of Priors9
This work proposes a novel Mixture of Priors (MoP) strategy, inspired by the Mixture-of-Experts (MoE) paradigm, that processes hyperprior information through multiple lightweight MLPs to generate diverse prior features, which are subsequently integrated into the MoP feature via a gating mechanism.
- Anatomical digital twins for medical education: a stepwise guide to create perpetual multimodal three-dimensional reconstruction of digital brain specimens9
A 3D digital model of human brain specimens is created using computed tomography and magnetic resonance imaging scans and combined with the powerful editing capabilities of the open-source 3D Slicer platform for image reconstruction and optimization to provide a digital 3D reconstruction model of the cranial brain specimen.
- 9
- 9
- Partitioned K-Means Clustering for Fast Construction of Unbiased Visual Vocabulary9
This work introduces a partitioned k-means clustering (PKM) scheme to efficiently generate a large and unbiased vocabulary using only a small training set and provides good scalability for building large vocabularies.
- 9
- 4DGS-CC: A Contextual Coding Framework for 4D Gaussian Splatting Data Compression8
This work introduces 4DGS-CC, a contextual coding framework that compresses 4DGS data to meet specific storage constraints and enables multi-rate 4DGS data compression tailored to specific storage requirements.
- 8
- Individual Content and Motion Dynamics Preserved Pruning for Video Diffusion Models7
A new Video Diffusion Model Compression approach using individual content and motion dynamics preserved pruning and consistency loss to gain comparable generation performance as larger VDM to VDMini, which significantly accelerates inference time while maintaining high-quality video generation.
- Proceedings of the first SIGMM workshop on Social media7
The program includes a keynote speech by Ramesh Jain on "Influence of Multimedia in Emerging Social Web Systems", and a panel discussion on "Emerging Challenges in Social Media Research".
- The reliability of radiologists' first impression interpreting a screening mammogram6
Although detecting gist signal is noisy, a sub-set of readers have the superior capability in detecting the gist of the abnormal and only the scores given by them are useful and reliable for predicting future breast cancer.
- 6
- 6
- High-resolution CT image retrieval using sparse convolutional neural network5
The proposed method has great potential in improving radiation dose calculation and delivery accuracy and decreasing CT radiation exposure of patients and the experimental results showed the proposed method could enhance images resolution from low-resolution images.
- 5
- A Flexible and Effective Linearization Method for Subspace Learning5
Linear discriminant analysis (LDA), as a supervised algorithm, aims to maximize the inter-class scatter and at the same time minimize the intra- class scatter.
- Progressive Classifier and Feature Extractor Adaptation for Unsupervised Domain Adaptation on Point Clouds4
This work proposes a novel framework that deeply couples the classifier and feature extractor adaption for 3D UDA, dubbed Progressive Classifier and Feature Extractor Adaptation (PCFEA).
- Unsupervised Part Discovery via Dual Representation Alignment4
The proposed method achieves competitive performance and robustness due to its part-specific attention and can further serve as a series of reliable detectors in the testing phase, predicting pixel masks for part discovery.
- 4
- IntersectGAN4
This paper proposes a novel GAN, namely IntersectGAN, to learn multiple attributes from different image domains through an intersecting architecture, and using the CelebFaces Attributes dataset, is able to produce high quality face images possessing multiple attributes.
- 4
- 4
- Understanding multimedia content using web scale social media data4
This three hour tutorial plans to introduce the important general concepts and themes of this timely topic, and review and summarize the recent multimedia content analysis methods using web-scale social media data as well as present insight into the challenges and future directions in this area.
- Robust global motion estimation method for aerial imagery4
A robust global motion model estimation method is proposed by fitting the optical flow field for aerial imagery by putting forward a simple optical flow valuation mechanism to choose a small set of reliable flows for fitting.
- PWFNet: Pyramidal Wavelet–Frequency Attention Network for Road Extraction3
This work proposes a novel frequency-aware road extraction network, termed PWFNet, which combines frequency-domain modeling with multi-scale feature enhancement and significantly improves road continuity, edge clarity, and robustness under complex conditions.
- Deep learning-based super-resolution US radiomics to differentiate testicular seminoma and non-seminoma: an international multicenter study3
It is demonstrated that a radiomics model of deep learning-based SR reconstructed US images enabled effective differentiation between SE and n-SE, enabling accurate, non-invasive preoperative differentiation between SE and n-SE.
- 3
- Introduction to the Special Issue on Multimodal Machine Learning for Human Behavior Analysis3
This special issue accepted nine articles that address the challenging issues of multimodal machine learning for human behavior analysis and focuses on modeling multiple features using sparse coding, while six deep learning–based methods have also been proposed to address some issues of multi-modality feature learning.
- 3
- 3
- Independent moving target detection for aerial video surveillance3
Without the need for image filtering, the independent moving targets can be efficiently detected by valuating extrema saliency with respect to different morphological attributes by an attribute based salientextrema detection algorithm.
- 4DGS-Craft: Consistent and Interactive 4D Gaussian Splatting Editing2
This work proposes 4DGS-Craft, a consistent and interactive 4DGS editing framework that can interpret user intent and decompose complex instructions into a logical sequence of atomic operations, enabling it to handle intricate user commands and further enhance editing performance.
- 2
- 2
- Guest Editorial: Special Issue on Deep Learning for Video Analysis and Compression2
Among twenty-three submissions, ten deep learning articles covering a broad range of topics on video analysis and compression were accepted, and these articles can be grouped into three categories: (1) deep learning for video analysis; (2) deepLearning for visual data processing and compression; and (3)Deep learning for visualData processing and compressor.
- 2
- Plane dimension detection with a single camera based on a flat refractive camera model2
A modified pinhole camera model is established based on the analysis of light refraction in the transparent plate placed between the scene and the camera, and an effective backward-projection algorithm is proposed to realise plane dimension detection and eliminate the plate refraction effect.
- Guest Editorial Special Section on Visual Saliency Computing and Learning2
The study on visual saliency models is mainly based on the biological evidence that humans’ visual attention is automatically attracted by highly salient features in the visual scene, which are discriminative with respect to the surrounding environment.
- Efficient Discriminative Learning of Class Hierarchy for Many Class Prediction2
This paper proposes a new efficient discriminative class hierarchy learning approach that can significantly accelerate the training process of the two state-of-the-art methods for multi-class tasks while achieving comparable prediction performance in terms of both classification accuracy and testing speed.
- Visual Content Identification and Search2
This special issue contains several research articles covering a diverse range of topics in visual content identification and search, concerning image and video content, by visual similarities.
- Proceedings of second ACM SIGMM workshop on Social media2
The program includes two keynote speeches, one by Ramesh Sarukkai on "Creating the Perfect Social Video Stream", and the other by Daniel Gatica- Perez on "The You in YouTube".
- 2
- 1
- 1
- 1
- [Relationship between morphological traits and body mass of Miichthys miiuy].1
The method of model fit test showed the accuracy of total length estimated body mass was higher than that of body height, and may provide an important reference for the effective use of morphological traits for breeding of M. miiuy.
- 1
- 1
- Proceedings of the 3rd ACM SIGMM international workshop on Social media1
The mission of this workshop is to give researchers and practitioners a unique opportunity to share their novel ideas, methods and techniques that address the emerging challenges of social media research, and to identify new directions for future research and development.
- FACE RECOGNITION — A GENERALIZED MARGINAL FISHER ANALYSIS APPROACH1
A new supervised learning algorithm to utilize the advantages of the Marginal Fisher Analysis and the Generalized Singular Value Decomposition techniques for face recognition, which demonstrates that GMFA outperforms LDA/Fisherface, L DA/GSVD and MFA.
- 1
- Metric Based Software Quality Assurance System1
A software quality assurance system named PMSQA1 is introduced in this paper and a multi-dimension process template library is designed in the sense that through simple tailoring, it can be applied directly to various types of organization.
- –
- From zero-shot to fine-tuning: optimize large language models for error detection of ultrasound reports–
It is demonstrated that task-specific fine-tuning enables open-source LLMs to rival proprietary LLMs and expert radiologists in Chinese ultrasound report error detection, offering a locally deployable and privacy-preserving, and regulation-compliant solution for enhancing reporting consistency and reducing diagnostic errors in high-volume ultrasound examination procedures.
- (RichMediaGAI'25) 3rd International Workshop on Rich Media with Generative AI–
The goal of this workshop is to showcase the latest advancements in generative AI for creating, editing, restoring, and compressing rich media data, including images, videos, and 3D content, using novel industry-level data.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Dynamic saliency detection via CNN and spatial-temporal fusion–
An approach to detect dynamic saliency of videos is proposed in this paper, which exploits a spatial-temporal fusion way and outperforms compared to four other dynamic Saliency models on two metrics.
- Theory and Experiment Study on Laser Self-Mixing Interference with Multiway Feedback External Cavity–
- –
- –
- –
- –
- –
- Automatic calibration of multi-camera networks–
This work developed an automatic calibration approach to calibrate the extrinsic parameters between different cameras that only needs to track the motion of objects in each camera from image sequences without the correspondence betweenDifferent cameras.
- –
- T-IRS–
This demonstration presents a (quasi) real-time textual query based image retrieval system (T-IRS) for consumer photos by leveraging millions of web images and their associated rich textual descriptions (captions, categories, etc.).
- –
- –
- –
- Divisional 3D shape reconstruction of object image–
A noel approach to reconstruct shape from shading information is presented, which considers both the global and local shading, and the distortion caused by image discontinuity but also the effects of local noises is eliminated.
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.