A5101886946- Retinal Imaging and Analysis
- Digital Imaging for Blood Diseases
- Tissue Engineering and Regenerative Medicine
- Electrospun Nanofibers in Biomedical Applications
- AI in cancer detection
- Tracheal and airway disorders
- Cutaneous Melanoma Detection and Management
- Radiomics and Machine Learning in Medical Imaging
- Nonmelanoma Skin Cancer Studies
- Glaucoma and retinal disorders
- Satellite Communication Systems
- Medical Image Segmentation Techniques
- Brain Tumor Detection and Classification
- Artificial Intelligence in Healthcare and Education
- Advanced Neural Network Applications
- Automated Road and Building Extraction
- Advanced Image Fusion Techniques
- IoT Networks and Protocols
- Industrial Vision Systems and Defect Detection
- Remote Sensing and LiDAR Applications
- Retinal Diseases and Treatments
- Magnetic Field Sensors Techniques
- Tactile and Sensory Interactions
- Power Line Inspection Robots
- Machine Learning and Data Classification
Sichuan University
2023-2025
University of Hong Kong
2025
Hubei University of Technology
2016-2024
Shandong Provincial Hospital
2024
Shandong First Medical University
2024
Yangzhou University
2021-2023
Wright State University
2023
Changsha University of Science and Technology
2023
Northern Jiangsu People's Hospital
2022
Hefei University of Technology
2021
The retinal vascular condition is a reliable biomarker of several ophthalmologic and cardiovascular diseases, so automatic vessel segmentation may be crucial to diagnose monitor them. However, existing methods have various problems in the vessels, such as insufficient weak anti-noise interference ability. Aiming shortcomings existed methods, this paper proposes an improved model based on Generative Adversarial Networks with U-Net, which contains densely-connected convolutional network novel...
Formation of functional and perfusable vascular network is critical to ensure the long-term survival functionality engineered tissue tracheae after transplantation. However, greatest challenge in tracheal-replacement therapy promotion regeneration by rapid graft vascularization. Traditional prevascularization methods for tracheal grafts typically utilize omentum or muscle flap wrapping, which requires a second operation; vascularized segment orthotopic transplantation one step remains...
Abstract The dense mechanoreceptors in human fingertips enable texture discrimination. Recent advances flexible electronics have created tactile sensors that effectively replicate slowly adapting (SA) and rapidly (RA) mechanoreceptors. However, the influence of dermatoglyphic structures on signal transmission, such as effect fingerprint ridge filtering friction‐induced vibration frequencies, remains unexplored. A novel multi‐layer sensor with an artificially synthesized skin surface capable...
Recently, the increasing deployment of LEO satellite systems has enabled various space analytics (e.g., crop and climate monitoring), which heavily relies on advancements in deep learning (DL). However, intermittent connectivity between satellites ground station (GS) significantly hinders timely transmission raw data to GS for centralized learning, while scaled-up DL models hamper distributed resource-constrained satellites. Though split (SL) can be a potential solution these problems by...
Functional vascularization is crucial for maintaining the long-term patency of tissue-engineered trachea and repairing defective trachea. Herein, we report construction evaluation a novel cell-free tracheal scaffold that effectively promotes graft. Our findings demonstrated exosomes derived from endothelial progenitor cells (EPC-Exos) enhance proliferation, migration, tube formation cells. Taking advantage angiogenic properties EPC-Exos, utilized methacrylate gelatin (GelMA) as carrier cell...
Abstract The segmentation of retinal vessels is greatly significant for doctors to diagnose the fundus diseases. However, existing methods have various problems in vessels, such as insufficient weak anti-noise interference ability. Aiming shortcomings existed methods, this paper proposes an improved model based on U-Net networks, which contains densely-connected convolutional network and a novel attention gate (AG) model, referred Densely-Attention-U-Net (DA-U-Net), automatically segment...
Long segment trachea defects are repaired by tracheal substitution, while decellularized technology has been effectively employed to prepare tissue engineering (TET).
BackgroundDeep learning methods are increasingly applied in the medical field; however, their lack of interpretability remains a challenge. Captum is tool that can be used to interpret neural network models by computing feature importance weights. Although an interpretable model, it rarely study problems, and there scarcity data regarding MRI anatomical measurements for patients with prostate cancer after undergoing Robotic-Assisted Radical Prostatectomy (RARP). Consequently, predictive...
Deep belief networks (DBNs) are graphical models which learn to extract the deep hierarchical representation of input data. Restricted boltzmann machines can be stacked and trained in a greedy manner form networks. The combination classifier is used for data classification. However, with increase size training set, it requires huge amount time get better classification accuracy results. To address these issues, novel fast classifier, called evolutionary function array voter (EFACV) proposed...
With the research background of rotorcraft flying robot's integrated navigation system,an error calibration method for electric magnetic compass is proposed.Based on analysis compass's measurement theory and sources,5 main factors which affect heading angle calculation precision compass,are summarized. In response to mentioned factors,software hardware solutions are proposed.Considering specialties robot,a simplified proposed calibrate hard magnetism errors scale factor errors.The can avoid...
This paper expounds method of the average weighted fusion, image pyramid wavelet transform and apply these methods on fusion multiple exposures nighttime images. Through calculating information entropy cross images, we can evaluate effect different fusion. Experiments showed that Laplacian algorithm is suitable for processing images it reduce halo while preserving details.
A de-noising approach is proposed that based on the combination of wiener filtering, nonlinear filtering and wavelet fusion, which de-noise LASCA (LAser Speckle Contrast Analysis) image blood vessels in Small Animals. The first performs laser spectral contrast analysis brain flow rats, get their spatial temporal images. Then, a method to deal with noise LASCA. restoration achieved by applying admixture subjective estimation objective are given As our experimental results shown, provides...
This paper presents a study on digitization of Traditional Chinese Medicine diagnosis. The research consists two aspects: a) algorithms for blood vessels extraction in sclera-conjunctiva images, which can be applied syndrome differentiation by observing human eyes (named Ocular Diagnostic Medicine); b) extracted vessels. First, region is isolated optimal threshold segmentation and mathematical operation; Scanning edge detection methods are used to gain the Moreover, feature parameters...
This paper presents 1) algorithms for blood vessels extraction in sclera-conjunctiva images, which can be applied syndrome differentiation by observing human eyes (named Ocular Diagnostic Traditional Chinese Medicine), 2) digitization of extracted vessels. First, region is isolated optimal threshold segmentation and mathematical operation. Second, scanning edge detection methods are used to gain the Moreover, feature parameters gained, reconstruct Experimental results show that obtain...
The explosive expansion of Artificial Intelligence over the past decade exhibits its potential as a technological podium for optimal decision-making infused by superintelligence, where human mind is restricted in ability to process enormous amounts data brief period. To accomplish predictive analysis or pattern recognition on massive set, machine learning way go. It most rapidly expanding path computer science, and considering health clinical informatics extremely obscure. objective ML...
As the demand for leather increases, automated defect identification is becoming increasingly important in industry. Based on issues that unobvious color of defects, small area and low recognition efficiency, this paper proposes a residual flow group attention network (ResFGANet) to effectively identify defects. First, convolution spatially separable are introduced feature extraction part compress model parameters improve efficiency recognition. Second, an improved efficient module added...