Computer aided endoscope diagnosis via weakly labeled data mining

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 机器人学研究室

	Computer aided endoscope diagnosis via weakly labeled data mining
	Wang S(王帅); Cong Y(丛杨); Fan HJ(范慧杰); Yang YS(杨云生); Tang YD(唐延东); Zhao HC(赵怀慈)
	2015
会议名称	2015 IEEE International Conference on Image Processing (ICIP)
会议日期	September 27-30, 2015
会议地点	Quebec City, QC, Canada
关键词	Computer aided diagnosis (CAD) multiple instance learning (MIL) weakly labeled endoscope images
页码	3072-3076
中文摘要	In comparison to most computer aided endoscope diagnosis methods using pixel-wise groundtruth by physicians manually, it is easy to get lots of endoscope images with corresponding diagnostic reports. In this paper, we intend to mine pixel-wise label information from these reports with weak frame-level labels automatically. To achieve this, we formulate our computer aided diagnosis problem as a Multiple Instance Learning (MIL) issue, where we represent each image as superpixels. Each image and each superpixel is cast as bag and instance, respectively. We then evaluate and select the most positive instances from positive bags automatically which helps us transform the frame-level classification problem into a standard supervised learning problem. In the experiment, we build a new gastroscopic image dataset with more than 3000 weakly labeled images, and ours outperforms the state-of-the-art methods, which verifies the effectiveness of our model.
收录类别	EI ; CPCI（ISTP）
产权排序	1
会议录	2015 IEEE International Conference on Image Processing (ICIP)
会议录出版者	IEEE
会议录出版地	Piscataway, NJ, USA
语种	英语
ISBN号	978-1-4799-8339-1
WOS记录号	WOS:000371977803040
内容类型	会议论文
源URL	[http://ir.sia.cn/handle/173321/17466]
专题	沈阳自动化研究所_机器人学研究室
推荐引用方式 GB/T 7714	Wang S,Cong Y,Fan HJ,et al. Computer aided endoscope diagnosis via weakly labeled data mining[C]. 见:2015 IEEE International Conference on Image Processing (ICIP). Quebec City, QC, Canada. September 27-30, 2015.