Uncertainty Quantification in Medical Image Segmentation

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 光电信息技术研究室

	Uncertainty Quantification in Medical Image Segmentation
	Li HX(李海星)1,2,3,4,5; Luo HB(罗海波)2,3,4,5
	2020
会议日期	December 11-14, 2020
会议地点	Chengdu, China
关键词	medical images uncertainty quantification segmentation prostate MRI image
页码	1936-1940
英文摘要	In medical images, the observer's manual description of different structures is very different, and it spans a wide range of various structures and pathologies. This variability (which is a characteristic of biological issues, imaging modality and expert annotators) has not been fully considered in the design of computer algorithms for medical image quantification. So far, few people predict the uncertainty of medical image segmentation. In this paper, we designed a U-shaped network to quantify the uncertainty in prostate MRI image segmentation. We have embedded a feature pyramid attention module in the backbone network, which can extract high-level semantic context information at different scales and provide a pixel-level attention to the decoder. At the same time, the module will not bring a large computational burden. In our experiments, we tested the performance of the proposed method on 55 clinical subjects.
产权排序	1
会议录	2020 IEEE 6th International Conference on Computer and Communications, ICCC 2020
会议录出版者	IEEE
会议录出版地	New York
语种	英语
ISBN号	978-1-7281-8635-1
内容类型	会议论文
源URL	[http://ir.sia.cn/handle/173321/29890]
专题	沈阳自动化研究所_光电信息技术研究室
通讯作者	Luo HB(罗海波)
作者单位	1.University of Chinese Academy of Sciences, Beijing, Chin 2.Institutes for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang Institute of Automation, Shenyang, China 3.Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, China 4.The Key Lab of Image Understanding and Computer Vision, Shenyan, China 5.Key Laboratory of Opto-Electronic Information Processing, Shenyang, Liaoning Province, China
推荐引用方式 GB/T 7714	Li HX,Luo HB. Uncertainty Quantification in Medical Image Segmentation[C]. 见:. Chengdu, China. December 11-14, 2020.