Yinwei Zhan

ORCID: 0000-0003-2899-9090
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Image Retrieval and Classification Techniques
  • Advanced Vision and Imaging
  • Image and Signal Denoising Methods
  • Advanced Image and Video Retrieval Techniques
  • Advanced Image Fusion Techniques
  • Medical Image Segmentation Techniques
  • Gaze Tracking and Assistive Technology
  • Hand Gesture Recognition Systems
  • Robotics and Sensor-Based Localization
  • Visual Attention and Saliency Detection
  • Video Surveillance and Tracking Methods
  • Advanced Neural Network Applications
  • Image Enhancement Techniques
  • Advanced Numerical Analysis Techniques
  • Human Pose and Action Recognition
  • Computer Graphics and Visualization Techniques
  • Image and Video Quality Assessment
  • Glaucoma and retinal disorders
  • Advanced Computing and Algorithms
  • Advanced Image Processing Techniques
  • AI in cancer detection
  • Muscle activation and electromyography studies
  • Image and Object Detection Techniques
  • Motor Control and Adaptation
  • EEG and Brain-Computer Interfaces

Guangdong University of Technology
2016-2025

Key Laboratory of Guangdong Province
2025

Sun Yat-sen University
2011-2012

Shantou University
1995-1996

Beijing Normal University
1994

To tackle the difficulties in detection and removal of impulse noise faced by existing filters, to further improve denoising performance, we propose an adaptive sequentially weighted median filter for image corrupted noise. In proposed method, a detector employing 3σ principle normal distribution local intensity statistics, is proposed; with neighborhood size, removal, which operator derived reference spatial distances from central noisy pixel, i.e., weighting coefficients are inversely...

10.1109/access.2019.2950348 article EN cc-by IEEE Access 2019-01-01

The recently proposed robust principle component analysis (RPCA) has been successfully applied in background subtraction. However, low-rank decomposition makes sense on the condition that foreground pixels (sparsity patterns) are uniformly located at scene, which is not realistic real-world applications. To overcome this limitation, we reconstruct input video frames and aim to make only sparse space but also time. Therefore, propose a joint frame set division RPCA-based method for In...

10.1109/tcsvt.2014.2333132 article EN IEEE Transactions on Circuits and Systems for Video Technology 2014-06-25

To overcome the drawbacks of existing filters for salt and pepper noises, an adaptive probability filter is proposed. For image, it detects noises based on characteristic minimum maximum intensity values images, as well distribution noise. If noise‐free intensities in neighbourhood repeat with a certain probability, highest repeated frequency used to remove noise statistical significance; otherwise, median pixels Experiments show that proposed method capable detecting more accurately perform...

10.1049/iet-ipr.2017.0910 article EN IET Image Processing 2018-01-04

ABSTRACT Magnetic resonance imaging (MRI) offers comprehensive information about brain structures, enabling excellent performance in tumor segmentation using multimodal MRI many methods. Nonetheless, missing modalities are common clinical practice, which can significantly degrade performance. Current methods often struggle to maintain feature consistency and robustness fusion when face difficulties accurately capturing boundaries. In this study, we propose an adaptive edge‐oriented...

10.1002/ima.70012 article EN International Journal of Imaging Systems and Technology 2025-01-01

ABSTRACT Motor learning involves complex interactions between the cognitive and sensorimotor systems, which are susceptible to different levels of task load. While mechanism underlying load‐dependent regulations in functions has been extensively investigated, their influence on downstream execution motor skill remains less understood. The current study extends understanding whether how alters activation pattern by a longitudinal functional near‐infrared spectroscopy (fNIRS) 30 healthy...

10.1002/hbm.70208 article EN cc-by-nc-nd Human Brain Mapping 2025-04-01

ABSTRACT Precise segmentation of pneumonia lesions using deep learning has been a research focus in medical image segmentation, which convolutional neural networks (CNNs) excel at capturing local features through layers but struggle with global information, while Transformers handle and long‐range dependencies well require substantial computational resources data. Motivated by the recently introduced Mamba that effectively models less complexity, we develop novel network architecture order...

10.1002/ima.70083 article EN International Journal of Imaging Systems and Technology 2025-04-10

The stereo-match to find the corresponding points in left and right images from an optical tracking system is a crucial step 3-D point reconstruction. Unfortunately, presence of ghost markers remains issue which caused by indistinct correspondence when several are coplanar with two centers. Such can lead erroneous medical instruments patients. Some methods have been proposed for this issue, but they rarely applied systems. To address we propose robust method combining maximum silhouette...

10.1109/access.2018.2869433 article EN cc-by-nc-nd IEEE Access 2018-01-01

This study investigated the complexity of electromyography (EMG) lower limb muscles when performing obstacle crossing tasks at different heights in poststroke subjects versus healthy controls. Five and eight controls were recruited to perform various (randomly set 10, 20, 30% leg's length). EMG signals recorded from bilateral biceps femoris (BF), rectus (RF), medial gastrocnemius, tibialis anterior during task. The fuzzy approximate entropy (fApEn) approach was used analyze signals. fApEn...

10.3389/fneur.2018.00131 article EN cc-by Frontiers in Neurology 2018-03-12

A method based on S transforms and support vector machines was presented for fault diagnosis of power electronics circuits in which the transform time-frequency analysis signal is used to extract features corresponding various faults. Then, types are identified through pattern recognition classifier, SVM. The simulation results show that proposed can accurately diagnose faults locate elements circuits. It also has excellent performance noise robustness calculation complexity, thus, having...

10.1016/j.scient.2011.06.013 article EN Scientia Iranica 2012-01-20

Due to the limitation of existing filters in detection and removal fixed value impulse noise, authors propose an iterative grouping median filter (IGMF) according characteristics noise intensity distribution. It sorts noise‐free pixels neighbourhood by intensity, divides sorted into groups depending on differences adjacent pixels, finally takes maximum group as noisy pixel. This strategy is performed iteratively full advantage previous denoising results. Experiments show that IGMF...

10.1049/iet-ipr.2018.6331 article EN IET Image Processing 2019-01-19

Patient-to-image registration is required for image-guided surgical navigation, but marker-based time consuming and subject to manual error. Markerless an alternative solution avoid these issues.This study designs a calibration board proposes geometric method calibrate the near-infrared tracking structured light components of proposed optical navigation system simultaneously.A planar cylinder are used evaluate accuracy calibration. The mean error experiment 0.035 mm, diameter 0.119 mm. A...

10.1002/rcs.1978 article EN International Journal of Medical Robotics and Computer Assisted Surgery 2018-12-17

Abstract MRI and CT images have been routinely used in clinical practice for treatment planning of the head‐and‐neck (HAN) radiotherapy. Delineating organs‐at‐risk (OAR) is an essential step radiotherapy, however, it time‐consuming prone to inter‐observer variation. The existing automatic segmentation approaches are either limited by image registration or lack global spatial awareness, thus under‐performed when dealing with complex anatomies. Herein, we propose a full‐scale attention network...

10.1049/ipr2.12663 article EN cc-by-nc IET Image Processing 2022-10-27

Eye tracking is a valuable topic in computer vision. Appearance-based eye promising research direction recent years. Convolutional neural networks (CNN) had been used gaze estimation, which cover the significant variability appearance caused by unconstrained head motion. With computation capability of consumer devices rapidly evolving, accurate and efficient appearance-based has potential for multipurpose applications. Person-independent have limit improving estimation accuracy....

10.1109/icaica50127.2020.9181854 article EN 2022 IEEE International Conference on Artificial Intelligence and Computer Applications (ICAICA) 2020-06-01

This paper presents a novel method for fingerprint image quality. Five features are extracted from the to analyze quality and feature vector is formed five features. Then SVM classifier, which can solve small-sample learning problems with good generalization, trained classify image. The separated into one of three classes, good-quality,medium-quality, or poor-quality. Experimental results on FVC2004 private database show that proposed an effective efficient scheme measure Our overcomes...

10.1109/paciia.2008.108 article EN 2008-12-01

Predicting gaze point on mobile devices without calibration in unconstrained environments has great significance human computer interaction. Appearance-based estimation methods have been improved due to the recent advance convolutional neural network (CNN) models and availability of large-scale datasets. CNN limitations extracting global information features ignore important local features. In this paper, we propose a novel structure named GazeAttentionNet. To improve accuracy estimation,...

10.1109/icassp43922.2022.9747911 article EN ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2022-04-27

Abstract Colonoscopy is one of the most direct and effective methods for detecting colon polyps; they are crucial early screening prevention colorectal cancer (CRC). Accurate segmentation polyp images significant clinical management treatment CRC. However, image a challenging task because polyps vary in size, shape, color, have low contrast with surrounding tissue mucosa. To address these challenges, we propose novel network called BGNet segmentation. consists three modules: boundary feature...

10.1002/ima.22959 article EN International Journal of Imaging Systems and Technology 2023-09-21
Coming Soon ...