A5041650010- Radiomics and Machine Learning in Medical Imaging
- Lung Cancer Diagnosis and Treatment
- Chronic Obstructive Pulmonary Disease (COPD) Research
- Medical Image Segmentation Techniques
- Retinal Imaging and Analysis
- AI in cancer detection
- Glaucoma and retinal disorders
- COVID-19 diagnosis using AI
- Respiratory Support and Mechanisms
- Medical Imaging Techniques and Applications
- 3D Shape Modeling and Analysis
- Retinal Diseases and Treatments
- Digital Imaging for Blood Diseases
- Image and Object Detection Techniques
- Image Retrieval and Classification Techniques
- Interstitial Lung Diseases and Idiopathic Pulmonary Fibrosis
- Colorectal Cancer Screening and Detection
- Computer Graphics and Visualization Techniques
- Esophageal Cancer Research and Treatment
- Retinal and Optic Conditions
- Digital Radiography and Breast Imaging
- Advanced Vision and Imaging
- Image Processing and 3D Reconstruction
- Inhalation and Respiratory Drug Delivery
- Advanced X-ray and CT Imaging
University of Pittsburgh
2016-2025
Xidian University
2024
UPMC Hillman Cancer Center
2023-2024
Bioengineering Center
2023
First Affiliated Hospital of Xi'an Jiaotong University
2014-2019
University of Pittsburgh Medical Center
2012-2019
Bioengineering (Switzerland)
2019
Peking University
2001-2017
Peking University Third Hospital
2017
University Radiology
2017
To develop and test computer software to detect, quantify, monitor progression of pneumonia associated with COVID-19 using chest CT scans. One hundred twenty scans from subjects lung infiltrates were used for training deep learning algorithms segment regions vessels. Seventy-two serial 24 detect quantify the presence COVID-19. The algorithm included (1) automated boundary vessel segmentation, (2) registration between scans, (3) computerized identification pneumonitis regions, (4) assessment...
Identification of pulmonary fissures, which form the boundaries between lobes in lungs, may be useful during clinical interpretation computed tomography (CT) examinations to assess early presence and characterization manifestation several lung diseases. Motivated by unique nature surface shape fissures 3-D space, we developed a new automated scheme using computational geometry methods detect segment depicted on CT images. After geometric modeling volume marching cubes algorithm, Laplacian...
To investigate the potential of deep learning in assessing pneumoconiosis depicted on digital chest radiographs and to compare its performance with certified radiologists.We retrospectively collected a dataset consisting 1881 X-ray images form radiography. These were acquired screening setting subjects who had history working an environment that exposed them harmful dust. Among these subjects, 923 diagnosed pneumoconiosis, 958 normal. identify we applied classical convolutional neural...
Low-dose CT (LDCT) is currently used in lung cancer screening of high-risk populations for early diagnosis. However, 96% individuals with detected nodules are false positives.In order to develop an efficient predictor from clinical, demographic and LDCT features, we studied a total 218 subjects or benign nodules. Probabilistic graphical models (PGMs) were integrate demographics, clinical data features 92 (training cohort) the Pittsburgh Lung Screening Study cohort.Learnt PGMs identified...
Purpose: To clarify whether and to what extent three-dimensional (3D) convolutional neural network (CNN) is superior 2D CNN when applied reduce false-positive nodule detections in the scenario of low-dose computed tomography (CT) lung cancer screening. Approach: We established a dataset consisting 1600 chest CT examinations acquired on different subjects from various sources. There were total 18,280 candidate nodules these examinations, among which 9185 9095 not nodules. For each nodule, we...
The study objective was to investigate if machine learning algorithms can predict whether a lung nodule is benign, adenocarcinoma, or its preinvasive subtype from computed tomography images alone.A dataset of chest scans containing nodules collected with their pathologic diagnosis several sources. split randomly into training (70%), internal validation (15%), and independent test sets (15%) at the patient level. Two were developed, trained, validated. first algorithm used support vector...
Lobe identification in computed tomography (CT) examinations is often an important consideration during the diagnostic process as well treatment planning because of their relative independence each other terms anatomy and function. In this paper, we present a new automated scheme for segmenting lung lobes depicted on 3-D CT examinations. The unique characteristic representation fissures form implicit functions using radial basis (RBFs), capable seamlessly interpolating ldquoholesrdquo...
In three-dimensional medical imaging, segmentation of specific anatomy structure is often a preprocessing step for computer-aided detection/diagnosis (CAD) purposes, and its performance has significant impact on diagnosis diseases as well objective quantitative assessment therapeutic efficacy. However, the existence various diseases, image noise or artifacts, individual anatomical variety generally impose challenge accurate structures. To address these problems, shape analysis strategy...