课程小组
¥
支付方式
请使用微信扫一扫 扫描二维码支付
请使用支付宝扫一扫 扫描二维码支付
分享主题
医学影像计算与分析
分享内容
人工智能应用于临床医学影像计算与分析成为近年来的研究热点,本次分享以2017年MICCAI录取论文为基础,选摘代表性的论文,介绍医学影像计算与分析的前沿方法和趋势。
《Suggestive Annotation: A Deep Active Learning Framework for Biomedical Image Segmentation》
论文地址:https://arxiv.org/abs/1706.04737
The authors present a deep active learning framework that combines fully convolutional network (FCN) and active learning to significantly reduce annotation effort by making judicious suggestions on the most effective annotation areas.
《Bayesian Image Quality Transfer with CNNs: Exploring Uncertainty in dMRI Super-Resolution》
论文地址:https://arxiv.org/pdf/1705.00664.pdf
In this work, the authors investigate the value of uncertainty modelling in 3D super-resolution with convolutional neural networks. A method was proposed to account for intrinsic uncertainty through a per-patch heteroscedastic noise model and for parameter uncertainty through approximate Bayesian inference in the form of variational dropout.《Automatic 3D Cardiovascular MR Segmentation with Densely-Connected Volumetric ConvNets》
论文地址:https://arxiv.org/abs/1708.00573
A DenseVoxNet with dense connectivity and sparing network architecture from a large number of redundant features is proposed to automatically segment the cardiac structures in the 3D cardiac MR images.
分享人
陈浩,视见医疗创始人兼首席科学家,在香港中文大学取得博士学位并获得香港政府博士奖学金,本科毕业于北京航空航天大学并获得金质奖章。研究兴趣包括医学影像计算,机器学习(深度学习), 计算机视觉等。博士期间发表数十篇顶级会议和期刊论文,包括CVPR、MICCAI、AAAI、MIA、IEEE-TMI、NeuroImage等 。担任包括NIPS、MICCAI、IEEE-TMI、NeuroImage等国际会议和期刊审稿人。三维全卷积神经网络相关论文获得2016 MIAR最佳论文奖。2014年以来带领团队在数十项国际性医学影像分析和识别挑战赛中获得冠军 。
分享时间
10月26日周四晚8点