A5066091439- Medical Imaging and Analysis
- Spine and Intervertebral Disc Pathology
- Medical Image Segmentation Techniques
- Spinal Fractures and Fixation Techniques
- Brain Tumor Detection and Classification
- Radiomics and Machine Learning in Medical Imaging
- Dental Radiography and Imaging
- Medical Imaging Techniques and Applications
- AI in cancer detection
- Neuroscience and Neuropharmacology Research
- Advanced X-ray and CT Imaging
- Sparse and Compressive Sensing Techniques
- Musculoskeletal pain and rehabilitation
- Aortic Disease and Treatment Approaches
- Topic Modeling
- Osteoarthritis Treatment and Mechanisms
- Radiation Dose and Imaging
- Cardiac Valve Diseases and Treatments
- Retinal Diseases and Treatments
- Retinal Imaging and Analysis
- Retinal and Optic Conditions
- Text and Document Classification Technologies
- Natural Language Processing Techniques
- Scoliosis diagnosis and treatment
- COVID-19 diagnosis using AI
Guangzhou Medical University
2022-2024
Southern Medical University
2015-2023
Spine parsing (i.e., multi-class segmentation of vertebrae and intervertebral discs (IVDs)) for volumetric magnetic resonance (MR) image plays a significant role in various spinal disease diagnoses treatments spine disorders, yet is still challenge due to the inter-class similarity intra-class variation images. Existing fully convolutional network based methods failed explicitly exploit dependencies between different structures. In this article, we propose novel two-stage framework named...
Aim Accurate severity grading of lumbar spine disease by magnetic resonance images (MRIs) plays an important role in selecting appropriate treatment for the disease. However, interpreting these complex MRIs is a repetitive and time-consuming workload clinicians, especially radiologists. Here, we aim to develop multi-task classification model based on artificial intelligence automated disc herniation (LDH), central canal stenosis (LCCS) nerve roots compression (LNRC) at axial MRIs. Methods...
Pulmonary hypertension (PH) is a fatal pulmonary vascular disease. The standard diagnosis of PH heavily relies on an invasive technique, i.e., right heart catheterization, which leads to delay in and serious consequences. Noninvasive approaches are crucial for detecting as early possible; however, it remains challenge, especially mild patients. To address this issue, we present new fully automated framework, hereinafter referred PHNet, noninvasively patients, improving the detection accuracy...
Hippocampus segmentation plays a significant role in mental disease diagnoses, such as Alzheimer's disease, epilepsy, and so on. Patch-based multi-atlas (PBMAS) approach is popular method for hippocampus has achieved promising result. However, the PBMAS needs high computation cost due to registration accuracy subject accuracy. In this paper, we propose novel based on iterative local linear mapping (ILLM) with representative structure-preserved feature embedding achieve accurate robust no...
BACKGROUND: Assessing the 3-dimensional (3D) relationship between critical anatomical structures and surgical channel can help select percutaneous endoscopic lumbar discectomy (PELD) approaches, especially at L5/S1 level. However, previous evaluation methods for PELD were mainly assessed using 2-dimensional (2D) medical images, making understanding of 3D lumbosacral difficult. Artificial intelligence based on automated magnetic resonance (MR) image segmentation has benefit reconstruction...
The three-dimensional (3D) anatomy of Kambin's triangle is crucial for surgical planning in minimally invasive spine surgery via the transforaminal approach. Few pieces research have, however, used image segmentation to explore 3D reconstruction triangle. This study aimed develop a new method based on automated magnetic resonance (MRI) lumbar spinal structures. An experienced (>5 years) "ground truth" pain physician meticulously segmented and labeled structures (e.g., bones, dura mater,...
Abstract We propose local linear mapping (LLM), a novel fusion framework for distance field (DF) to perform automatic hippocampus segmentation. A k-means cluster method is constructing magnetic resonance (MR) and DF dictionaries. In LLM, we assume that the MR samples are located on two nonlinear manifolds from manifold differentiable locally linear. combine dictionary using representation present test sample, corresponding coefficients derived procedure predict of sample. then merge...
The lumen of aortic dissection (AD) has important clinical value for preoperative diagnosis, interoperative intervention, and post-operative evaluation AD diseases. segmentation is challenging because (i) fitting its irregular profile by using traditional models difficult, (ii) the size image usually so big that many algorithms have to perform down-sampling reduce computational burden, thereby reducing resolution result. In this paper, an automatic algorithm, in which a 3D mesh gradually...
(1) Background: This study aims to develop a deep learning model based on 3D Deeplab V3+ network automatically segment multiple structures from magnetic resonance (MR) images at the L4/5 level. (2) Methods: After data preprocessing, modified of was used for automatic segmentation MR We performed five-fold cross-validation evaluate performance model. Subsequently, Dice Similarity Coefficient (DSC), precision, and recall were also assess model's performance. Pearson's correlation coefficient...
Supervised deep-learning techniques with paired training datasets have been widely studied for low-dose computed tomography (LDCT) imaging excellent performance. However, the are usually difficult to obtain in clinical routine, which restricts wide adoption of supervised practices. To address this issue, a general idea is construct pseudo dataset based on available unpaired data, after which, can be adopted improving LDCT performance by dataset. due complexity noise properties CT imaging,...
Automatic estimation of axial spine indices is clinically desired for various computer aided procedures, such as disease diagnosis, therapeutic evaluation, pathophysiological understanding, risk assessment, and biomechanical modeling. Currently, the are manually measured by physicians, which time-consuming laborious. Even worse, tedious manual procedure might result in inaccurate measurement. To deal with this problem, paper, we aim at developing an automatic method to estimate multiple from...
Automated retinal vessel segmentation is crucial to the early diagnosis and treatment of ophthalmological diseases. Many deep-learning-based methods have shown exceptional success in this task. However, current approaches are still inadequate challenging vessels (e.g., thin vessels) rarely focus on connectivity segmentation.We propose using an error discrimination network (D) distinguish whether pixel predictions (S) correct, S trained obtain fewer D. Our method similar to, but not same as,...
Automatic estimation of indices from medical images is the main goal computer-aided quantification (CADq), which speeds up diagnosis and lightens workload radiologists. Deep learning technique a good choice for implementing CADq. Usually, to acquire high-accuracy quantification, specific network architecture needs be designed given CADq task. In this study, considering that target organs are intervertebral disc dural sac, we propose an object-specific bi-path (OSBP-Net) axial spine image...
Automated multi-class segmentation of vertebrae and intervertebral discs for volumetric magnetic resonance image(MRI) can help the diagnosis treatment many spinal diseases. However, existing methods based on fully convolutional network mostly stacked local convolution pooling operations, thus failed to capture long-range dependencies spine segmentation. In this paper, we propose a novel computation-efficient residual U-Net with strip-pooling attention mechanism (SPA-ResUNet) discs. Due...
3D reconstruction of lumbar intervertebral foramen (LIVF) has been beneficial in evaluating surgical trajectory. Still, the current methods reconstructing LIVF model are mainly based on manual segmentation, which is laborious and time-consuming. This study aims to explore feasibility automatically segmenting spinal structures increasing speed accuracy magnetic resonance image (MRI) at L4-5 level.A total 100 participants (mean age: 42.2 ± 14.0 years; 52 males 48 females; mean body mass index,...
Automated measurement of spine indices on axial magnetic resonance (MR) images plays a significant role in lumbar spinal stenosis diagnosis. Existing direct approaches fail to explicitly focus the task-specific region or feature channel with additional information for guiding. We aim achieve accurate by introducing guidance segmentation task.In this paper, we propose segmentation-guided regression network (SGRNet) automated measurement. SGRNet consists path generating prediction and...
Low-dose computed tomography (CT) is of great potential advantage for disease diagnosis. Usually, paired training datasets are difficult to obtain in clinical routine, which catalyzes the development unsupervised learning techniques improve low-dose CT imaging. Recently, most existing approaches imaging were developed image domain, and only a few have been sinogram challenging task. In this paper, we propose dedicated unpaired technique restoration with novel data-dependent noise-generative...
Deformation field estimation is an important and challenging issue in many medical image registration applications. In recent years, deep learning technique has become a promising approach for simplifying problems, been gradually applied to registration. However, most existing registrations do not consider the problem that when receptive cannot cover corresponding features moving fixed image, it output accurate displacement values. fact, due limitation of field, 3 x kernel difficulty...