Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks70 from public data
- 39
- NTIRE 2025 Challenge on Day and Night Raindrop Removal for Dual-Focused Images: Methods and Results37
This paper reviews the NTIRE 2025 Challenge on Day and Night Raindrop Removal for Dual-Focused Images to establish a new and powerful benchmark for the task of removing raindrops under varying lighting and focus conditions.
- 37
- 37
- Unified Data-Free Compression: Pruning and Quantization without Fine-Tuning36
This paper proposes a novel framework named Unified Data-Free Compression (UDFC), which performs pruning and quantization simultaneously without any data and fine-tuning process, and achieves significant improvements over various network architectures and compression methods.
- Selective Learning Confusion Class for Text-Based CAPTCHA Recognition36
A novel method of selective learning confusion class for text-based CAPTCHA recognition with a view on improving the recognition accuracy of the confusing characters in confusion-class DCNN is presented.
- A Learning Framework for n-Bit Quantized Neural Networks Toward FPGAs35
To solve the gradient vanishing problem, this article proposes a reconstructed gradient function for QNNs in the back-propagation algorithm that can directly get the real gradient rather than estimating an approximate gradient of the expected loss.
- 33
- NTIRE 2025 Challenge on Low Light Image Enhancement: Methods and Results29
This paper thoroughly evaluates the state-of-the-art advancements in LLIE, showcasing the significant progress.
- Low-Light Image Enhancement via Generative Perceptual Priors24
A novel LLIE framework with the guidance of Generative Perceptual Priors (GPP-LLIE) derived from vision-language models (VLMs) is presented and it is demonstrated that this model outperforms current SOTA methods on paired LL datasets and exhibits superior generalization on real-world data.
- 24
- Ntire 2025 Image Shadow Removal Challenge Report22
This work examines the findings of the NTIRE 2025 Shadow Removal Challenge, with 17 teams successfully submitting their solutions during the final evaluation phase, and one focusing on reconstruction fidelity and the other on visual perception through a user study.
- 22
- Deep RAW Image Super-Resolution. A NTIRE 2024 Challenge Survey21
The NTIRE 2024 RAW Image Super-Resolution Challenge is reviewed, highlighting the proposed solutions and results, and providing a gauge for the current state-of-the-art in RAW Image Super-Resolution.
- 20
- Research on improved algorithm of object detection based on feature pyramid20
Compared with SSD, the accuracy of FP-SSD is higher, which has more accurate location and higher recognition confidence, and the algorithm applied the deconvolution network to the high-level features of the feature pyramid to get the semantic information and convolution for the middle level features to reduce the feature channels.
- 3QNet18
A novel learning-based point cloud compression framework named 3D Point Cloud Geometry Quantiation Compression Network (3QNet) is proposed, which overcomes the robustness limitation of existing point-based methods and can handle dense points.
- Learning Global-aware Kernel for Image Harmonization16
A novel Global-aware Kernel Net-work (GKNet) is proposed to harmonize local regions with comprehensive consideration of long-distance background references for local harmonization and a novel Selective Correlation Fusion (SCF) module is proposed to better select relevant long-distance background references for local harmonization.
- 16
- 16
- 14
- 13
- 13
- 13
- Out-of-Stock Detection Based on Deep Learning13
An out-of-stock detection method based on deep learning based on canny operator performs better, which can achieve a recall rate of 83.9%, a precision rate of 91.7% and a F1-Measure rate of 87.6% in real scene dataset.
- Diffusion-Based Imaginative Coordination for Bimanual Manipulation12
A multi-frame latent prediction strategy that encodes future states in a compressed latent space, preserving task-relevant features and a unidirectional attention mechanism where video prediction is conditioned on the action, while action prediction remains independent of video prediction is introduced.
- 12
- 11
- Enhanced Semantic Information Transfer of Multi-Domain Samples: An Adversarial Edge Detection Method Using Few High-Resolution Remote Sensing Images11
A weakly supervised domain adaptation method to address the high dependence of edge extraction networks on samples and prevents deviations in the semantic features that are caused by the overgeneralization of edge features is proposed.
- 10
- 9
- 8
- Propagating Asymptotic-Estimated Gradients for Low Bitwidth Quantized Neural Networks8
A novel Asymptotic-Quantized Estimator (AQE) to estimate the gradient of loss function through the graph flow with a derivative of 0 almost everywhere is proposed and a M-bit Inputs and N-bit Weights Network (MINW-Net) trained by AQE, a quantized neural network with 1–3 bits weights and activations.
- 8
- Wavelet Theory and Application Summarizing8
The wavelet analysis is the development based on Fourier transform and it is a breakthrough of the Fourier analysis, and wavelet transform is particularly good at dealing with non-stationary signal analysis.
- Retinex-Guided Histogram Transformer for Mask-Free Shadow Removal7
ReHiT is proposed, an efficient mask-free shadow removal framework based on a hybrid CNN-Transformer architecture guided by Retinex theory that delivers competitive results with one of the smallest parameter sizes and fastest inference speeds among top-ranked entries, highlighting its applicability for real-world applications with limited computational resources.
- Learning to Train a Point Cloud Reconstruction Network Without Matching6
This work proposes a novel framework named PCLossNet which learns to train a point cloud reconstruction network without any matching, and proves the superiority of this method, where PCLossNet can help networks achieve much lower reconstruction errors and extract more representative features.
- 6
- Resolution-Free Point Cloud Sampling Network with Data Distillation5
A novel resolution-free point clouds sampling network to directly sample the original point cloud to different resolutions is proposed, conducted by optimizing non-learning-based initial sampled points to better positions by optimizing non-learning-based initial sampled points to better positions.
- 5
- 5
- 4
- 4
- 3
- 3
- Extraction and Portrait of Knowledge Points for Open Learning Resources2
This article explores how to use the technology of text summarization and keyword extraction to automatically extract key knowledge points from massive educational resources and use open resources to generate feature portraits of relevant knowledge points.
- Improved zerotree coding algorithm for wavelet image compression2
A listless minimum zerotree coding algorithm based on the fast lifting wavelet transform with lower memory requirement and higher compression performance is presented in this paper and a comparison reveals that the PSNR results of LMZC are higher than those of LZC, and the compression performance ofLMZC outperforms that of SPIHT in terms of hard implementation.
- 1
- A model developed by machine learning and principal component analysis for predicting bronchiolitis obliterans syndrome after allogeneic HSCT in a Chinese population with hematologic malignancies1
The MLR model, developed through data-driven machine learning, effectively identifies high-risk BOS populations at an early stage and emerged as pivotal in predicting BOS risk post-allo-HSCT.
- OOPS: Outlier-aware and quadratic programming based structured pruning for large language models1
This work proposes a structured pruning framework named OOPS (Outlier-aware and quadratic prOgramming based Structured Pruning), which comprises three key components: outlier-aware pruning unit selection, quadratic programming based reconstruction, and layer-wise distillation.
- Hyperbolic Binary Neural Network1
This article transforms the constrained problem in hyperbolic space into an unconstrained one in Euclidean space using the Riemannian exponential map and proposes the exponential parametrization cluster (EPC) method, which, compared with the Riemannian exponential map, shrinks the segment domain based on a diffeomorphism.
- Pse: mixed quantization framework of neural networks for efficient deployment1
This work proposes PSE, a mixed quantization framework which combines product quantization (PQ), scalar quantization (SQ), and error correction, and demonstrates that PSE reduces computation complexity and model size with acceptable accuracy loss.
- 1
- Agent-Based User Model for Personalized Retrieval System1
Information retrieval system combined with Agent technology has the autonomy, responsiveness, cooperation, learning and other characteristics, to construct an Agent-based personalized search system user model, to provide users with a personalized search service.
- 1
- A Novel Discriminant Non-Negative Matrix Factorization and its Application to Facial Expression Recognition1
The experiments on facial expression recognition indicate that the algorithm enhances the discrimination capability of low-dimensional features and achieves better performance than other non-negative matrix factorization algorithms.
- 1
- WLR: Well-conditioned linear reconstruction for retraining-free pruning of LLMs–
The proposed Well-conditioned Linear Reconstruction (WLR), a structured pruning compensation method that avoids introducing extra parameters into the pruned model, reconstructs pruned layers using a linear combination of preserved channels while addressing the issue of ill-conditioning during the process.
- –
- –
- –
- –
- –
- –
- –
- Tracking an Object over 200 FPS with the Fusion of Prior Probability and Kalman Filter–
A novel robust approach fusing the prediction information of Kalman filter and prior probability is proposed for tracking arbitrary objects and it achieves over 200 fps on the Intel i3 CPU for tracking an arbitrary object on benchmark.
- –
- –
- A Design of High Performance Simulation Platform–
This paper designs a high proficiency simulation platform and develops related technical structure, so as to provide a full-cycle of modeling and simulation (M&S) to support complex system.
- –
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.