学术报告：Single-channel Sparse Nonnegative Blind Source Separation Method for Automatic 3D Delineation of Lung Tumors in PET Images

报告时间：星期二（2017.4.18）下午2：00

报告地点：电子信息楼306

报告人： Ivica Kopriva, Ph.D, senior scientist, the Ru?erBoškovi? Institute, Zagreb, Croatia

邀请人： 陈新建 特聘教授

欢迎广大师生参加！

Abstract

Recently, novel method for automatic segmentation of lung tumor from 3D PET images was proposed by us (I. Kopriva, et al. IEEE Journal of Biomedical and Health Informatics, doi:10.1109/JBHI.2016.2624798). In particular, we combine feature expansion transform with sparseness constrained nonnegative matrix factorization to decompose 3D PET image into constituent components. Afterwards, by using complexity based criterion, we select tumor component as the one with minimal complexity. The proposed method was compared with threshold based on 40% and 50% maximum standardized uptake value (SUV), graph cuts (GC), random walks (RW) and affinity propagation (AP) algorithms on 18 non-small cell lung cancer datasets with respect to ground truth provided by two radiologists. Dice similarity coefficient averaged with respect to two ground truths is: 0.78±0.12 by the proposed algorithm, 0.78±0.1 by GC, 0.77±0.13 by AP, 0.77±0.07 by RW, and 0.75±0.13 by 50% maximum SUV threshold. Since the proposed method achieved performance comparable with interactive methods, considering the unique challenges of lung tumor segmentation from PET images, our findings support possibility of using our fully automated method in routine clinics.

Biography

Ivica Kopriva obtained PhD degree from the Faculty of Electrical Engineering and Computing, University of Zagreb in 1998 with a subject in blind source separation. From 2001 till 2005 he was research and senior research scientist at Department of Electrical and Computer Engineering, The George Washington University, Washington D.C., USA. Since 2006 he is senior scientist at the Ru?erBoškovi? Institute, Zagreb, Croatia. His research interests are related to development of algorithms for unsupervised learning with applications in biomedical image analysis, chemometrics and bioinformatics. He published over 40 papers in internationally recognized journals and holds 3 US patents. He is co-author of the research monograph: Kernel Based Algorithms for Mining Huge Data Sets: Supervised, Semi-supervised and Unsupervised Learning, Springer Series: Studies in Computational Intelligence, 2006. He is senior member of the IEEE and the OSA.