您现在的位置:网站首页>通知公告

通知公告

【医学部学术讲座】——ROI Contrast Enhancement and CNNs Adaptation to Medical Image Applications

文章来源: 作者: 发布时间:2017年06月29日 点击数: 字体:

主讲嘉宾: Minh Hung Le 华中科技大学 博士

时间: 2017年7月6日上午 10:00 – 11:30

地点: 深圳大学医学院附楼学术报告厅

主持人: 雷柏英 讲师

主讲嘉宾简介:

Minh Hung Le received the B.Eng. degree from the Ho Chi Minh City University of Transport, Vietnam, in 2008, and the M.Eng. and Ph.D. degrees from Huazhong University of Science and Technology, Wuhan, China, in 2013 and 2017 respectively, all in electronics and information engineering. During his Ph.D. program, he was working under the supervision of Prof. Wenyu Liu and Prof. Xin Yang in the area of Medical Image Analysis. His Ph.D. dissertation mainly focuses on designing new enhancement methods for robustly and accurately extracting regions of interests in medical images. In particular, he developed the maximally stable temporal volumes which could effectively enhance the kidney in dynamic contrast enhancement MRI images. This technique is further applied to renal compartment segmentation and renal function analysis. This work was collaborated with Prof. Kyung Hyun Sung (UCLA) and was originally published in MICCAI 2015 conference, a top-tier conference for medical image analysis.

报告内容简介:

Processing medical image is a challenging task due to the inhomogeneity of image contrast, noise, image artifacts, etc. There are numerous dedicated algorithms and methods have been proposed for medical image analysis and achieved high performance in many medical application tasks. Those research efforts provide valuable tools that can aid the radiologists for establishing more accurate diagnosis and appropriate treatment planning. Two image processing and analysis techniques applied in three specific medical image applications which recently attract many research interests are presented in this talk including region of interests (ROIs) contrast enhancement and CNNs adaptation to Medical Image Applications. ROIs contrast enhancement aims at enhancing the contrast at each specific ROI of organs or tissues and help to reduce the manual setting of thresholding parameters for the segmentation of each specific ROI. In deep convolutional neural networks (CNNs) adaptation to the application of Prostate Cancer (PCa) diagnosis in multi-parametric magnetic resonance images (MP-MRI), the proposed multimodal CNN can effectively and efficiently extract and fuse image features from multiple sources of MP-MRI image modalities.

【打印文章】 【添加收藏】