Institute of
Bioinspired Intelligence and Mining Knowledge
Location:Home > News > 中国科学技术大学周文罡教授学术报告会


2018-09-20 11:34:45

报告题目:Image Retrievalwith Deep Learning




报告摘要:Recentyears has witnessed the great success of deep learning in a variety of visiontasks. In most cases, deep learning is conducted in a supervised way. As forimage search, since the category number of potential objects is difficult toenumerate and the image database is large, it is infeasible to collectsufficient annotated training images as supervision for deep learning. As a result,most works on image search simply leverage the activations from pre-traineddeep learning model, or just focus on some specific fine-grained tasks, such aslandmark retrieval. To this end, we explore deep learning in apseudo-supervised paradigm and orient it for image retrieval. We approach itfrom different perspectives and propose three algorithms. Further, toautomatically evaluate the retrieval result quality, we propose a deep learningbased quality assessment method. Extensive experiments demonstrate the effectiveness and potential of pseudo-supervised deep learning in retrieval task.

报告人简介:周文罡,中国科学技术大学信息学院副教授、博士生导师,国家优秀青年基金获得者。2006年本科毕业于武汉大学,2011年博士毕业于中国科学技术大学,2011~2013年在美国德州大学圣安东尼奥分校计算机系任博士后研究员,2013年加入中国科学技术大学,任职至今。主要研究方向为多媒体信息检索和计算机视觉,发表IEEETPAMIIEEETIPTMMIEEE汇刊论文和CVPRACM MMCCF A类国际会议论文共40余篇,获国际会议ICIMCS 2012最佳论文奖。Google学术引用2400余次,SCI他引600余次,ESI高被引论文两篇。曾获中科院百篇优博奖,入选中科院青促会会员、中国科协"青年人才托举工程"