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
- Tien‐Tsin WongSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works119 from public data
- R³Net: Recurrent Residual Refinement Network for Saliency Detection522
A novel recurrent residual refinement network (R^3Net) equipped with residual refinement blocks (RRBs) to more accurately detect salient regions of an input image that outperforms competitors in all the benchmark datasets.
- SINet: A Scale-Insensitive Convolutional Neural Network for Fast Vehicle Detection341
A scale-insensitive convolutional neural network (SINet) for fast detecting vehicles with a large variance of scales is presented and a context-aware RoI pooling to maintain the contextual information and original structure of small scale objects is presented.
- Bidirectional Feature Pyramid Network with Recurrent Attention Residual Modules for Shadow Detection285
This paper presents a network to detect shadows by exploring and combining global context in deep layers and local context in shallow layers of a deep convolutional neural network (CNN) and develops a bidirectional feature pyramid network (BFPN) to aggregate shadow contexts spanned across different CNN layers.
- Learning Semantic Context from Normal Samples for Unsupervised Anomaly Detection241
This work presents a Semantic Context based Anomaly Detection Network, SCADN, for unsupervised anomaly detection by learning the semantic context from the normal samples by generating multi-scale striped masks to remove a part of regions from thenormal samples and training a generative adversarial network to reconstruct the unseen regions.
- Deep Texture-Aware Features for Camouflaged Object Detection165
Experimental results show that the approach outperforms various state-of-the-art methods by a large margin, and the network is evaluated on the benchmark datasets for camouflaged object detection both qualitatively and quantitatively.
- 138
- Deep Attentional Features for Prostate Segmentation in Ultrasound123
A novel deep neural network equipped with deep attentional feature (DAF) modules for better prostate segmentation in TRUS by fully exploiting the complementary information encoded in different layers of the convolutional neural network (CNN).
- Hybrid dual Kalman filtering model for short‐term traffic flow forecasting112
A novel hybrid dual Kalman filter is presented for accurate and timely short-term traffic flow forecasting and outperforms the state-of-the-art parametric and non-parametric models.
- δ -agree AdaBoost stacked autoencoder for short-term traffic flow forecasting102
An adaptive boosting scheme is proposed to ensemble the trained stacked autoencoders to improve the accuracy of traffic flow forecasting and sufficient experiments have been conducted to demonstrate the superior performance of the proposal.
- 68
- Structure-based ASCII art68
This paper presents a novel method to generate structure-based ASCII art that approximates the major line structure of the reference image content with the shape of characters using a novel alignment-insensitive shape similarity (AISS) metric.
- A Learning-Based Multimodel Integrated Framework for Dynamic Traffic Flow Forecasting66
A novel deep learning-based multimodel integration framework in order to overcome the limitations of previous methods in dealing with large variations and uncertainties of traffic flow and hence improve the forecasting accuracy.
- SVRGSA: a hybrid learning based model for short‐term traffic flow forecasting64
This study proposes a hybrid traffic flow forecasting model combining gravitational search algorithm (GSA) and the SVR model, whereby the GSA is employed to search optimal SVR parameters.
- Context-Aware and Scale-Insensitive Temporal Repetition Counting63
A context-aware and scale-insensitive framework, to tackle the challenges in repetition counting caused by the unknown and diverse cycle-lengths, and proposes a bidirectional cycle length estimation method for a context- aware prediction.
- Animating animal motion from still62
This paper infer motion cycle of animals from the "motion snapshots" (snapshots of different individuals) captured in a still picture to demonstrate the applicability of the proposed method to a wide range of species, including birds, fishes, mammals, and reptiles.
- 55
- 52
- 52
- G²Face: High-Fidelity Reversible Face Anonymization via Generative and Geometric Priors42
G2Face is introduced, which leverages both generative and geometric priors to enhance identity manipulation, achieving high-quality reversible face anonymization without compromising data utility.
- 42
- From Continuity to Editability: Inverting GANs with Consecutive Images41
This paper introduces consecutive images into the inversion process, providing the first support of video-based GAN inversion and an interesting application of unsupervised semantic transfer from consecutive images.
- 37
- Rolling normal filtering for point clouds36
Compared with other state-of-the-art denoising methods, this approach is more robust in removing small-scale geometric features while retaining large-scale structures and exhibits superiority in preventing large- scale sharp structures from severe distortion.
- 32
- D4Net: De-deformation defect detection network for non-rigid products with large patterns28
A D4Net is proposed to detect defects of a non-rigid product with deformation in a given image and its corresponding reference image by focusing on differences between high-level semantic features extracted from the deep neural network to emphasize the region of possible defects.
- Highlight-assisted nighttime vehicle detection using a multi-level fusion network and label hierarchy28
A feature aggregation mechanism to combine multi-scale highlight features and the vehicle’s visual features to take advantages of the vehicle highlight position hints is proposed and the performance of the method goes beyond that of the state-of-the-art.
- 27
- Where is My Spot? Few-shot Image Generation via Latent Subspace Optimization26
This paper addresses this dilemma by projecting sparse few-shot samples into a continuous latent space that can potentially generate infinite unseen samples, and injects a novel class into the StyleGAN latent subspace, and therefore new unseen samples can be easily produced by sampling images from it.
- 25
- Towards a Smaller Student: Capacity Dynamic Distillation for Efficient Image Retrieval23
A Capacity Dynamic Distillation framework is proposed, which constructs a student model with editable representation capacity, initially a heavy model to fruitfully learn distilled knowledge in the early training epochs, and the stu-dent model is gradually compressed during the training.
- 23
- 23
- Text-aware balloon extraction from manga23
This paper proposes an automatic approach that can extract text balloons from manga images both accurately and effectively and incorporates text properties in order to differentiate text blobs from texture blobs.
- 22
- 21
- D3still: Decoupled Differential Distillation for Asymmetric Image Retrieval18
The Decoupled Differential Distillation (D3still) framework shifts from absolute one-to-one supervision to optimizing the relational differences in pairwise similarities produced by the query and gallery networks, thereby preserving a consistent retrieval order across both networks.
- Pose- and Attribute-consistent Person Image Synthesis18
A Pose- and Attribute-consistent Person Image Synthesis Network (PAC-GAN) is proposed, which renders target appearances by explicitly transferring the appearance information from the source image to the target image in a component-wise manner and eliminates source-target matching ambiguity.
- Dual pyramid network for salient object detection18
This work presents a dual pyramid network (DPNet) for salient object detection by formulating self-attention mechanism into the sub-regions based contexts by designing the spatial pyramid module and channel-wise pyramid module.
- 17
- 17
- 17
- 16
- DALocNet: Improving Localization Accuracy for Domain Adaptive Object Detection16
This paper presents a novel deep neural network design for domain adaptive object detection by refining the pseudo labels generated from the current object detection methods and using these labels with a weighted loss function to train the network on the target domain.
- Beyond Textual Constraints: Learning Novel Diffusion Conditions with Fewer Examples15
This paper delves into a novel aspect of learning novel diffusion conditions with datasets an order of magnitude smaller, and implements condition-specific negative rectification, which addresses the inconsistencies typically brought about by Classifier-free guidance in few-shot training con-texts.
- 14
- ViewSRD: 3D Visual Grounding Via Structured Multi-View Decomposition14
ViewSRD is a framework that formulates 3D visual grounding as a structured multi-view decomposition process that significantly outperforms state-of-the-art methods, particularly in complex queries requiring precise spatial differentiation.
- 14
- 14
- 13
- RecDreamer: Consistent Text-to-3D Generation via Uniform Score Distillation13
This work proposes a solution called RecDreamer, which reshapes the underlying data distribution to achieve a more consistent pose representation, and effectively mitigates the Multi-Face Janus problem, leading to more consistent 3D asset generation across different poses.
- 13
- GDFace: Gated Deformation for Multi-View Face Image Synthesis13
This paper proposes a Gated Deformable Face Synthesis Network to model the deformation of faces that aids the synthesis of the target face image, and introduces a soft-gating mechanism that enables adaptive fusion between deformable features and primitive features.
- CIRI: Curricular Inactivation for Residue-aware One-shot Video Inpainting12
This paper resolves the one-shot video inpainting problem in which only one annotated first frame is provided and proposes curricular inactivation to replace the hard masking mechanism for indicating the in-painting target, which is robust to erroneous predictions in long-term video inpainting.
- Object Detection in Densely Packed Scenes via Semi-Supervised Learning with Dual Consistency12
Experimental results corroborate that the proposed method is able to improve the object detection performance in densely packed scenes, consistently outperforming state-of-the-art approaches.
- Rotation-Adaptive Point Cloud Domain Generalization via Intricate Orientation Learning11
This work identifies the most challenging rotation for each point cloud and construct an intricate orientation set by optimizing intricate orientations, and employs an orientation-aware contrastive learning framework that incorporates an orientation consistency loss and a margin separation loss, enabling effective learning of categorically discriminative and generalizable features with rotation consistency.
- Learning an Interpretable Stylized Subspace for 3D-Aware Animatable Artforms11
The rationale is to transfer the interpretability and 3D controllability of the latent space in a 3D-aware GAN to a stylized sub-space of a customized GAN, revitalizing the original artforms, enabling 3D-aware animatable reconstruction and editing.
- 11
- AHE Detection With a Hybrid Intelligence Model in Smart Healthcare11
A novel methodology combining with deep network and multiple-gene expression programming is presented to predict the AHE automatically which will improve the efficiency of prevention and lighten the burden of diagnosis for ICU doctors.
- Illumination-Invariant and Deformation-Tolerant Inner Knuckle Print Recognition Using Portable Devices11
This work is the first to simultaneously consider the illumination invariance and deformation tolerance for appearance-based low-resolution hand biometrics and proposes a novel structure-context descriptor based on local statistics.
- 10
- Study on Bezier Curve Variable Step-length Algorithm10
By changing the parameter step-length of the curve generation algorithm, it can significantly reduce the calculations of a large number of duplicate points in the point-by-point generation algorithm; the algorithm not only maintains a higher accuracy, but also significantly improves the efficiency generated in the curve, with better application.
- Adaptive Selecting and Learning Network and a New Benchmark for Imbalanced Fine-Grained Ship Classification9
A new framework called Adaptive Selecting and Learning Network (ASL) is designed, to solve the problem of fine-grained classification of ship natural images in the real world, and outperforms most of the existing fine- grained classification methods.
- Region-based structure line detection for cartoons9
This paper presents the first approach to extract structure lines in cartoons based on regions, which starts by segmenting an image into regions, and then classifies them as edge regions and non-edge regions.
- 8
- Appearance-Preserved Portrait-to-Anime Translation via Proxy-Guided Domain Adaptation8
This article starts from learning the mapping between coser and anime portraits, and presents a proxy-guided domain adaptation learning scheme with three progressive adaptation stages to shift the initial model to the human portrait domain, which can generate visually pleasant anime portraits with well-preserved appearances given the human portrait.
- 8
- 8
- Texture-Aware ASCII Art Synthesis with Proportional Fonts8
A novel feature extraction scheme based on a multi-orientation phase congruency model is developed, which aims at faithfully capturing visually sensitive features, including both main contours and textural structures, while suppressing visually insensitive features.
- StarPose: 3D Human Pose Estimation via Spatial-Temporal Autoregressive Diffusion7
This paper proposes StarPose, an autoregressive diffusion framework that effectively incorporates historical 3D pose predictions and spatial-temporal physical guidance to significantly enhance both the accuracy and temporal coherence of pose predictions.
- VrdONE: One-stage Video Visual Relation Detection7
This work proposes VrdONE, a streamlined yet efficacious one-stage model that achieves state-of-the-art performances on the VidOR benchmark and ImageNet-VidVRD, showcasing its superior capability in discerning relations across different temporal scales.
- Towards High-Quality Visualization of Superfluid Vortices7
Novel techniques to aid the visual understanding of superfluid vortices based on the state-of-the-art non-linear Klein-Gordon equation are presented, which evolves a complex scalar field giving rise to special vortex lattice/ring structures with dynamic vortex formation, reconnection, and Kelvin waves, etc.
- GenPoly: Learning Generalized and Tessellated Shape Priors via 3D Polymorphic Evolving6
GenPoly, a novel generalized 3D prior model designed for multiple 3D generation tasks, focuses on preserving fine details, and formulating 3D prior modeling as a bottom-up polymorphic evolving process, which consistently surpasses state-of-the-art methods across various downstream tasks.
- 6
- 6
- 6
- 5
- Seeing 3D Through 2D Lenses: 3D Few-Shot Class-Incremental Learning via Cross-Modal Geometric Rectification5
This work proposes Cross-Modal Geometric Rectification (CMGR), a framework that enhances 3D geometric fidelity by leveraging CLIP's hierarchical spatial semantics and introduces a Structure-Aware Geometric Rectification module that hierarchically aligns 3D part structures with CLIP's intermediate spatial priors through attention-driven geometric fusion.
- 5
- 4
- 4
- 4
- 4
- 3
- Unambiguous granularity distillation for asymmetric image retrieval3
GranDist constructs layered feature representations that balance the richness of contextual information with the granularity of local features that outperforms state-of-the-art techniques and significantly enhances the performance of previous asymmetric retrieval approaches.
- 3
- FR²Seg: Continual Segmentation Across Multiple Sites via Fourier Style Replay and Adaptive Consistency Regularization3
This paper introduces FR2Seg, to sufficiently exploit both domain-specific and domain-invariant knowledge for efficient continual learning with the aid of low-frequency cues and presents a Fourier style replay module to synthesize pseudo images with old-site styles for data augmentation during new-site training, effectively preventing catastrophic forgetting without sacrificing data privacy.
- 3
- Think2Sing: Orchestrating Structured Motion Subtitles for Singing-Driven 3D Head Animation3
This work proposes Think2Sing, a unified diffusion-based framework that integrates pretrained large language models to generate semantically consistent and temporally coherent 3D head animations conditioned on both lyrics and acoustics and constructs the first multimodal singing dataset with synchronized 3D motion, acoustic descriptors, and aligned motion subtitles, enabling semantically grounded and expressive motion learning.
- A Two-Phase Space Resection Model for Accurate Topographic Reconstruction from Lunar Imagery with PushbroomScanners3
Experimental results evidence that the proposed mathematical model can obtain more accurate EOPs and topographic maps not only for the simulated data, but also for the real data from Chang’E-1, compared to the existing space resection model.
- T2SGrid: Temporal-to-Spatial Gridification for Video Temporal Grounding2
The core idea of T2SGrid is to process video content in clips rather than individual frames, which enables the use of composite text timestamps to establish global temporal awareness and achieves superior performance on standard VTG benchmarks.
- 2
- L3Net: Localized and Layered Reparameterization for incremental learning2
This work proposes a novel three-stage CIL framework called L3Net, which introduces a Localized Dual-path Expansion structure, which allows the model to learn simultaneously from both old and new features by integrating a fusion selector after each convolutional layer.
- SCJD: Sparse Correlation and Joint Distillation for Efficient 3D Human Pose Estimation2
Sparse Correlation and Joint Distillation (SCJD) is proposed, a novel framework that balances efficiency and accuracy for 3D HPE and introduces Sparse Correlation Input Sequence Downsampling to reduce redundancy in student network inputs while preserving inter-frame correlations.
- 2
- StableGuard: Towards Unified Copyright Protection and Tamper Localization in Latent Diffusion Models1
- Tgcpn: two-level grid context propagation network for 3D small object detection1
A new 3D backbone, named Two-level Grid Context Propagation Network (TGCPN), which integrates two grid representations into a single framework, which exhibits excellent performance in 3D small object detection.
- 1
- 1
- Face Expression Recognition via Product-Cross Dual Attention and Neutral-Aware Anchor Loss1
A neutral-expression-aware expression feature similarity loss based on the traditional anchor loss, which can further guide the network to learn better features from an input image, and which can better leverage the landmarks to extract more discriminative features from an input image.
- BroadCAM: Outcome-agnostic Class Activation Mapping for Small-scale Weakly Supervised Applications1
Since broad learning system (BLS) is independent to the model learning, BroadCAM can avoid the weights being affected by the unreliable model outcomes when with small-scale data, and achieves SOTA performance with large-scale training data.
- 1
- Fast scene labeling via structural inference1
This paper proposes a bi-directional recurrent network to model the information flow along the parent-child path via structural inference in a fast LSTM scene labeling network, and it outperforms state-of-the-art methods on different benchmarks.
- 1
- Regular Super-resolution Restoration Algorithm Based on Fuzzy Similarity Fusion for a Single Haze Image over the Ocean1
Experimental results show that the proposed algorithm can effectively improve the restoration effect of haze images over the sea.
- 1
- 1
- –
- RecPoint: Zero-Shot Text-to-3D Point Cloud Generation Via Structure-Aware Local Semantic Rectification–
RecPoint is introduced, a structure-aware framework for text-guided 3D shape generation that rectifies semantic drift by explicitly modeling local geometric relationships and proposes a test-time locality-aware semantic rectification mechanism that bridges modality gaps by grounding textual features in visually similar, structurally aligned image embeddings.
- Large model-assisted video summarization via global entity unification and robust importance scoring–
A summarization pipeline around Global Entity Unification and Robust Importance Scoring is built, but unlike earlier efforts, each object is traced across frames and attached consistent identifiers to it, and Fragmented, isolated descriptions become a single, object-aware text corpus that unifies the storyline.
- –
- NavOne: One-Step Global Planning for Vision-Language Navigation on Top-Down Maps–
This work proposes Top-Down VLN (TD-VLN), reformulating navigation as a one-step global path planning problem on pre-built top-down maps, supported by the newly constructed R2R-TopDown dataset.
- Invert4TVG: A Temporal Video Grounding Framework with Inversion Tasks Preserving Action Understanding Ability–
By inverting TVG to derive query-related actions from segments, this approach strengthens semantic understanding, significantly raising the ceiling of localization accuracy.
- –
- –
- VolumeSwap: Volumetric Decomposed 3D‐Aware Face Swapping–
This work adopts a novel volumetric perspective, deconstructing the 3D‐aware face swapping problem into two distinct volumes: the inner swapped face and the background, and introduces VolumeSwap, featuring dual branches specifically tailored for each volume.
- –
- –
- Erratum to “Multi-View Face Synthesis via Progressive Face Flow”–
In the above article [1], unfortunately, Fig. 5 was not displayed correctly with many empty images, so the correct version is supplemented here.
- Collective Intelligence for Electric Power Knowledge System–
A personalized recommendation method joins weighted slope one algorithm and clustering algorithm, which will improve the calculation accuracy and reduce the computational complexity of the electric power knowledge database.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.