A5100634161- Advanced Neural Network Applications
- Retinal Imaging and Analysis
- Medical Image Segmentation Techniques
- Medical Imaging and Analysis
- Ophthalmology and Visual Impairment Studies
- Ocular Diseases and Behçet’s Syndrome
- Advanced Computational Techniques and Applications
- Glaucoma and retinal disorders
- Image Processing Techniques and Applications
- Intraocular Surgery and Lenses
- Orthopedic Infections and Treatments
- COVID-19 diagnosis using AI
- Multimodal Machine Learning Applications
- Rough Sets and Fuzzy Logic
- Multimedia Communication and Technology
- AI in cancer detection
- Airway Management and Intubation Techniques
- Video Coding and Compression Technologies
- Digital Media Forensic Detection
- Non-Invasive Vital Sign Monitoring
- Image Retrieval and Classification Techniques
- Fiber-reinforced polymer composites
- Graphene research and applications
- Retinal and Optic Conditions
- Dental Radiography and Imaging
Tsinghua University
2021-2024
China State Construction Engineering (China)
2024
China University of Geosciences
2024
Sun Yat-sen University
2020-2023
Xidian University
2023
Sichuan University
2023
Shanghai Sixth People's Hospital
2022
Shanghai Jiao Tong University
2022
Xijing University
2020
Shanghai University of Engineering Science
2007
Automatic segmentation of liver tumors is crucial to assist radiologists in clinical diagnosis. While various deep learningbased algorithms have been proposed, such as U-Net and its variants, the inability explicitly model long-range dependencies CNN limits extraction complex tumor features. Some researchers applied Transformer-based 3D networks analyze medical images. However, previous methods focus on modeling local information (eg. edge) or global morphology) with fixed network weights....
The convolutional neural network has achieved remarkable results in most medical image seg- mentation applications. However, the intrinsic locality of convolution operation limitations modeling long-range dependency. Although Transformer designed for sequence-to-sequence global prediction was born to solve this problem, it may lead limited positioning capability due insufficient low-level detail features. Moreover, features have rich fine-grained information, which greatly impacts edge...
Background: Myopia is the leading cause of visual impairment and affects millions children worldwide. Timely annual manual optometric screenings entire at-risk population improve outcomes, but screening challenging due to lack availability training assessors economic burden imposed by screenings. Recently, deep learning computer vision have shown powerful potential for disease screening. However, these techniques not been applied large-scale myopia using ocular appearance images. Methods: We...
To assess associations of high academic performance with ametropia prevalence and myopia development in Chinese schoolchildren.This multicohort observational study was performed Guangdong, China. We first a cross-sectional cohort analysis students grades 1 to 9 from Yangjiang evaluate the relationship between refractive status on yearly basis. also longitudinal analyses Shenzhen trend changes over period 33 months. All statuses were measured using noncycloplegic autorefractors.A total 32,360...
Deformable image registration is widely utilized in medical analysis, but most proposed methods fail the situation of complex deformations. In this paper, we present a cascaded feature warping network to perform coarse-to-fine registration. To achieve this, shared-weights encoder adopted generate pyramids for unaligned images. The module then used estimate deformation field at each level. manner implemented by cascading from bottom level top Furthermore, multi-scale loss also introduced...
Infantile cataract is the main cause of infant blindness worldwide. Although previous studies developed artificial intelligence (AI) diagnostic systems for detecting infantile cataracts in a single center, its generalizability not ideal because complicated noises and heterogeneity multicenter slit-lamp images, which impedes application these AI real-world clinics. In this study, we two lens partition strategies (LPSs) based on deep learning Faster R-CNN Hough transform improving detection. A...
Distal femur fractures are complex injuries with a high rate of fracture healing problems. Since the widespread computed tomographic imaging in diagnosis distal fractures, many characteristics have been discovered. This study aimed to depict location and frequency lines further analyze morphological using 3-dimensional tomography (CT) mapping technique, thus providing more information solve this challenging clinical problem.In total, 217 216 patients were retrospectively reviewed. Fracture...
Prosthetic joint infection (PJI) is a prevalent and severe complication characterized by high diagnostic challenges. Currently, unified standard incorporating both computed tomography (CT) images numerical text data for PJI remains unestablished, owing to the substantial noise in CT disparity volume between data. This study introduces method, HGT, based on deep learning multimodal techniques. It effectively merges features from scan patients’ via Unidirectional Selective Attention (USA)...
Recent advances in multimodal imaging acquisition techniques have allowed us to measure different aspects of brain structure and function. Multimodal fusion, such as linked independent component analysis (LICA), is popularly used integrate complementary information. However, it has suffered from missing data, commonly occurring neuroimaging data. Therefore, this paper, we propose a Full Information LICA algorithm (FI-LICA) handle the data problem during fusion under framework. Built upon...
We present a linear, arbitrary ratio resizer for the interoperability of digital video in MPEG applications. The supports conversions among interlaced/progressive, 4:2:0, 4:2:2 and 4:4:4 formats with resizing fractional shift. Particularly, interlaced to progressive conversion (deinterlacing), combined deinterlacing algorithm is proposed, which bandwidth efficient hardware implementation. With proposed algorithm, filters can be generated implemented on-the-fly using software. also useful...
Visual impairment is a widespread public health issue that can have negative impact on quality of life, education and socioeconomic development.1 The early stages particularly childhood (infancy toddlerhood), are crucial periods for visual development, during which detection treatment ocular pathology prevent irreversible vision loss.2, 3 Unfortunately, young children often unable to complain symptoms or unwilling participate in standard tests, making it challenging assess their functions....
Prosthetic Joint Infection (PJI) is a prevalent and severe complication characterized by high diagnostic challenges. Currently, unified standard incorporating both computed tomography (CT) images numerical text data for PJI remains unestablished, owing to the substantial noise in CT disparity volume between data. This study introduces method, HGT, based on deep learning multimodal techniques. It effectively merges features from scan patients' via Unidirectional Selective Attention (USA)...
To improve the detection accuracy of Lugus Lucorum in natural environment, a method is proposed based on improved convolution neural network (CNN). First, YOLO-v3, new training data set labeling strategy designed to make target have higher effective pixel occupation rate Ground Truth. Two DenseBlock structures are integrated effectively alleviate gradient disappearance, reduce number parameters, and save computational power. Feature reuse can also play an anti-overfitting role. The validated...
Abstract Background: Lens opacity seriously affects the visual development of infants. Slit-illumination images play an irreplaceable role in lens detection; however, these exhibited varied phenotypes with severe heterogeneity and complexity, particularly among pediatric cataracts. Therefore, it is urgently needed to explore effective computer-aided method automatically diagnose heterogeneous provide appropriate treatment recommendations a timely manner. Methods: We integrated three...
Deformable image registration is widely utilized in medical analysis, but most proposed methods fail the situation of complex deformations. In this paper, we pre-sent a cascaded feature warping network to perform coarse-to-fine registration. To achieve this, shared-weights encoder adopted generate pyramids for unaligned images. The module then used estimate deformation field at each level. manner implemented by cascading from bottom level top Furthermore, multi-scale loss also introduced...