Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks94 from public data
- Each Part Matters: Local Patterns Facilitate Cross-View Geo-Localization329
It is argued that neighbor areas can be leveraged as auxiliary information, enriching discriminative clues for geo-localization, and introduced a simple and effective deep neural network, called Local Pattern Network (LPN), to take advantage of contextual information in an end-to-end manner.
- Infrared Small Target Detection with Scale and Location Sensitivity292
A novel Scale and Location Sensitive (SLS) loss is proposed to handle the limitations of existing losses and a simple Multi-Scale Head to the plain U-Net (MSHNet) is designed, which outperforms existing state-of-the-art methods by a large margin.
- Age-Invariant Face Recognition by Multi-Feature Fusionand Decomposition with Self-attention238
A novel Multi-feature Fusion and Decomposition framework for age-invariant face recognition, which learns more discriminative and robust features and reduces the intra-class variants and improves the AIFR performance than previous state-of-the-art methods.
- 152
- Image Demoireing with Learnable Bandpass Filters126
Experimental results on two public datasets show that the novel multiscale bandpass convolutional neural network (MBCNN) outperforms state-of-the-art methods by a large margin.
- Implicit Dual-Domain Convolutional Network for Robust Color Image Compression Artifact Reduction106
An implicit dual-domain convolutional network with a pixel position labeling map and quantization tables as inputs is proposed and is superior to the state-of-the-art methods and IDCN-f exhibits excellent abilities to handle a wide range of compression qualities with a little trade-off against performance.
- 95
- Toward Understanding and Boosting Adversarial Transferability From a Distribution Perspective89
This work proposes a novel method that crafts adversarial examples by manipulating the distribution of the image and can significantly improve the transferability of the crafted attacks and achieves state-of-the-art performance in both untargeted and targeted scenarios.
- 87
- SDPL: Shifting-Dense Partition Learning for UAV-View Geo-Localization78
A dense partition strategy (DPS) is proposed, dividing the image into multiple parts to explore contextual information while explicitly maintaining the global structure, and the shifting-fusion strategy, which generates multiple sets of parts in parallel based on various segmentation centers.
- 77
- 65
- AIM 2019 Challenge on Image Demoireing: Methods and Results57
This paper reviews the first-ever image demoireing challenge that was part of the Advances in Image Manipulation (AIM) workshop, held in conjunction with ICCV 2019, and focuses on the proposed solutions and their results.
- 50
- NTIRE 2020 Challenge on Image Demoireing: Methods and Results45
This paper reviews the Challenge on Image Demoireing that was part of the New Trends in Image Restoration and Enhancement (NTIRE) workshop, held in conjunction with CVPR 2020, and examines the current state-of-the-art in image and burst image demoireing problems.
- Improving Dynamic HDR Imaging with Fusion Transformer42
This paper proposes a transformer model for HDR imaging that outperforms the state-of-the-art by large margins on several popular public datasets, and demonstrates that the key component is the HDR transformer module.
- 41
- 31
- Information-Containing Adversarial Perturbation for Combating Facial Manipulation Systems29
This work proposes a novel two-tier protection method named Information-containing Adversarial Perturbation (IAP), which provides more comprehensive protection for facial images and introduces a feature-level correlation measurement that is more suitable to measure the difference between the facial images than the commonly used mean squared error.
- 29
- Bidirectional difference locating and semantic consistency reasoning for change captioning28
This work proposes a framework with bidirectional difference localization and semantic consistency reasoning to describe the image changes and shows that the performance of the proposed model outperforms the state‐of‐the‐art change captioning models by a large margin.
- 26
- S-Net: a scalable convolutional neural network for JPEG compression artifact reduction25
Experimental results indicate that the proposed S-Net CNN outperforms other CNN-based methods and achieves state-of-the-art performance.
- 20
- Quad Bayer Joint Demosaicing and Denoising Based on Dual Encoder Network with Joint Residual Learning19
A novel dual encoder network, namely DRNet, to achieve joint demosaicing and denoising for Quad Bayer CFA and shows that the proposed method could enhance the reconstruction quality from the native ISP algorithm.
- DomainPlus: Cross Transform Domain Learning towards High Dynamic Range Imaging17
This paper proposes a cross-transform domain neural network for efficient HDR imaging, and proposes fronted Discrete Wavelet Transform and Discrete Cosine Transform-based learnable bandpass filters to formulate aCross- transform domain learning block, dubbed DomainPlus Block (DPB) for effective ghosting removal.
- Non-local degradation modeling for spatially adaptive single image super-resolution12
It is demonstrated that SASR can efficiently estimate degradation representations and handle multiple types of degradation, and leads to state-of-the-art blind SISR performance when compared to alternative approaches.
- 11
- SMTPD: A New Benchmark for Temporal Prediction of Social Media Popularity10
It is verified that temporal alignment and early popularity play crucial roles in social media popularity prediction for not only deepening the understanding of temporal dynamics of popularity in social media but also offering a suggestion about developing more effective prediction models in this field.
- 10
- 10
- Rethinking Out-of-Distribution Detection From a Human-Centric Perspective10
A human-centric evaluation of OOD detection is proposed and it is found that the simple baseline OOD detection method can achieve comparable and even better performance than the recently proposed methods, which means that the development in OOD detection in the past years may be overestimated.
- Aerial-view geo-localization based on multi-layer local pattern cross-attention network10
This work proposes a Transformer-based network, named Multi-layer Local Pattern Cross Attention Network (MLPCAN), which employs the cross-attention block (CAB) to establish correlations between information of feature maps from different layers when images are fed into the network.
- 9
- 9
- 8
- 8
- 7
- Progressive Inertial Poser: Progressive Real-Time Kinematic Chain Estimation for 3-D Full-Body Pose From Three IMU Sensors6
This work proposes a method called progressive inertial poser (ProgIP) for human pose estimation, which combines neural network estimation with a human dynamics model, considers the hierarchical structure of the kinematic chain, and employs a multistage progressive network estimation with increased depth to reconstruct full-body motion in real time.
- 6
- 6
- Depth-guided deep filtering network for efficient single image bokeh rendering6
A depth-guided deep filtering network, called DDFN, is proposed in this study, and the focused region detection block is designed to detect the salient areas, and the depth estimated block is introduced to estimate depth maps from full-focus images.
- Rich Structural Index for Stereoscopic Image Quality Assessment6
A Rich Structural Index (RSI) for Stereoscopic Image objective Quality Assessment (SIQA) method based on multi-scale perception characteristics has stable performance and strong competitive advantage.
- DepthDark: Robust Monocular Depth Estimation for Low-Light Environments5
This work introduces a flare-simulation module and a noise-simulation module to accurately simulate the imaging process under nighttime conditions, producing high-quality paired depth datasets for low-light conditions and presents an effective low-light PEFT strategy that utilizes illumination guidance and multiscale feature fusion to enhance the model's capability in low-light environments.
- Capturing Stable HDR Videos Using a Dual-Camera System5
An asynchronous dual-camera system (DCS) is designed, which enables independent exposure control across two cameras, eliminating the need for synchronization typically required in traditional multi-camera setups, and an exposure-adaptive fusion network (EAFNet) is formulated for the DCS system.
- 5
- Multi-stage reasoning on introspecting and revising bias for visual question answering5
A bias discrimination module is proposed to distinguish good and bad bias indiscriminately in existing debiasing methods and helps with feature updating for both image regions and question words, thereby facilitating more accurate and comprehensive reasoning about both the image content and question.
- DCT-DiffPose: A Lightweight Diffusion Model With Multi-Hypothesis for 3D Human Pose Estimation4
DCT-DiffPose, a novel framework that integrates a diffusion model with Confidence and Consistency-based Multi-Hypothesis Aggregation (CCMA) and incorporate the Discrete Cosine Transform (DCT) for frequency-domain feature extraction, demonstrates its excellent trade-off between accuracy and complexity.
- 4
- Deeper feature integration network for salient object detection of strip steel surface defects4
Comparison results show that the proposed saliency model performs better than the 15 state-of-the-art methods, which proves the effectiveness and superiority of the proposed model.
- Gain and Phase Calibration of Uniform Rectangular Arrays Based on Convex Optimization and Neural Networks4
A calibration method based on convex optimization (CVX) and neural networks is proposed for the large planar arrays of phased array three-dimensional imaging sonar systems and the efficiency of the proposed method for gain and phase calibration is demonstrated.
- Hard parameter sharing for compressing dense-connection-based image restoration network4
The distribution of convolution layers is studied and a hard parameter sharing approach known as convolution pool (CP) is proposed for compressing dense-connection-based image restoration CNN models, which can efficiently achieve an impressive compression rate with negligible performance reduction.
- Evolution of ICTs-empowered-identification: A general re-ranking method for person re-identification4
A Feature Relation Map to mine the latent relation between the k -neighbors through convolution neural network is introduced, and a metric learning based Similarity Evaluation (SE) model is proposed to obtain the re-ranking distance from the FRM by incorporating the dilated convolution and normal convolution kernels.
- WXSOD: A Benchmark for Robust Salient Object Detection in Adverse Weather Conditions3
This paper introduces a novel Weather-eXtended Salient Object Detection (WXSOD) dataset, and proposes an efficient baseline, termed Weather-aware Feature Aggregation Network (WFANet), which adopts a fully supervised two-branch architecture.
- Counterfactual GAN for debiased text-to-image synthesis3
A new debiased text-to-image synthesis method with a counterfactual learning scheme that conceptualizes the process as calculating the Natural Direct Effect on the generated visual features and use the Total Indirect Effect to produce the final image for debiased text-to-image generation.
- 3
- 3
- 3
- A Unified Asymmetric Knowledge Distillation Framework for Image Classification3
Experimental results show that the proposed concise distillation framework can achieve considerable improvements over existing methods, demonstrating the effectiveness of the method and the potential of the framework.
- 3
- 2
- 2
- Build receptive pyramid for efficient color image compression artifact reduction2
This work presents an RP-based convolution neural network named receptive pyramid convolutional network (RPCN) for efficient color image compression artifact reduction (CAR) with a much smaller model and much faster running speed for the CAR task.
- 1
- 1
- 1
- Gaussian bare‑bone JAYA algorithm for multi-threshold medical image segmentation1
The proposed GBJAYA algorithm incorporates Gaussian-distributed random number update mechanisms to enhance global search capabilities and accelerate convergence, offering superior performance, stability, and fast convergence in medical image segmentation.
- 1
- Reference-aware image harmonization1
A Style-aware Reference Module that utilizes a diffusion model as an estimator to generate reference style features, thereby achieving global style alignment and a Region-aware Reference Module that adaptively identifies and utilizes relevant reference regions between the foreground and background by predicting soft masks.
- SympCoughNet: symptom assisted audio-based COVID-19 detection1
This work proposes SympCoughNet, a deep learning-based COVID-19 audio classification network that integrates cough sounds with clinical symptom data and demonstrates significant performance improvements over traditional audio-only approaches.
- EHPE: A Segmented Architecture for Enhanced Hand Pose Estimation1
A novel segmented architecture for enhanced hand pose estimation (EHPE) is proposed, which performs a local extraction of the TIP and wrist, thus alleviating the effect of error accumulation on the prediction of the TIP and reducing the predictive errors for all joints on this basis.
- Hierarchical Frequency-based Upsampling and Refining for Compressed Video Quality Enhancement1
This work proposes a hierarchical frequency-based upsampling and refining neural network (HFUR) for compressed video quality enhancement that achieves state-of-the-art performance for both constant bit rate and constant QP modes.
- 1
- 1
- 1
- –
- –
- Depth map super resolution grounded in the multi-dimensional collaborative guidance–
A novel framework that combines the efficiency of State Space Models (SSM) with multi-dimensional guidance, which exploits the linear complexity of SSMs to model long-range dependencies, ensuring structurally consistent depth recovery across large smooth regions without the heavy burden of self-attention is proposed.
- –
- –
- –
- –
- Perceptual Quality Assessment of Real-World Noise in Ultra-High-Definition Images–
PINE is proposed, a novel noise-oriented BIQA method that achieves state-of-the-art performance on the UNIQ-5k dataset and high rank consistency with PSNR and SSIM across the evaluated SIDD denoising results, showing that it may serve as an auxiliary no-reference indicator for restoration-oriented comparisons.
- –
- Low-light Image Enhancement with Retinex Decomposition in Latent Space–
A Retinex-Guided Transformer~(RGT) model, which is a two-stage model consisting of decomposition and enhancement phases, to separate reflectance and illumination components and convert the intrinsically multiplicative relationship into an additive formulation, enhancing decomposition stability and precision.
- Enhancing out-of-distribution detection with bilateral distribution score–
This work proposes the Bilateral Distribution Score (BDS), which effectively integrates both OOD and ID scores to enhance detection performance and maintains backward compatibility with existing ID scoring techniques, requiring no architectural modifications or retraining procedures.
- Position-Sensitive painterly image harmonization–
A Position-Sensitive Painterly Image Harmonization method that integrates both global and local background styles in a position-sensitive way, ensuring the inserted object harmonizes with both the overall style and localized variations is proposed.
- –
- –
- –
- –
- –
- –
- 3D Indoor Scene Completion via Room Layout Estimation–
The task of layout-based hole filling for the incomplete indoor scene scans is proposed: from the mesh of a scene model, the scene layout is estimated by detecting the principal planes of ascene and leverage the layout as the prior for the accurate completion of planar regions.
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.