题 目: Deep hashing for large-scale image and video retrieval
主 讲: 王瑞平 副研究员
时 间：2017年6月15日（周四）下午 14：00
地 点：数据科学与jbo竞博电竞官方网站A101学术报告厅
主 持：郑伟诗
报告摘要:Recent years have witnessed the explosive growth of image and video data on the Internet, posing great challenges to retrieving images/videos relevant to a given query image/video. At the meantime, the retrieval tasks have also become more diverse, such as 1) retrieving images from the same category, 2) retrieving images with specified attributes, and 3) the combination of the above tasks, e.g. looking for clothing of the same style but with a different color. To deal with such tasks, hashing is often adopted for its high efficiency in both time and storage. In this talk, I will introduce recent progresses in our group towards this topic. First, to tackle the traditional category retrieval, we propose a novel Deep Supervised Hashing (DSH) method that takes pairs of images as training inputs and learns the desired compact binary codes as the output of each image in an end-to-end manner. A loss function is elaborately designed to maximize the discriminability of the output space. Then, to address the aforementioned multiple retrieval tasks, we propose a unified framework named Dual Purpose Hashing (DPH), which jointly preserves the category and attribute similarities by exploiting CNN networks in a multi-task learning fashion. Our loss functions are designed to make use of the abundant partially labelled data on the Internet, which can meanwhile improve the generalization ability of the models. Furthemore, we extend the idea of DSH to Deep Video Code (DVC) that learns a single binary code for each video clip to solve the task of face video retrieval.
主讲人简介:王瑞平，于2003年获北京交通大学理学学士学位，2010年获中科院计算所工学博士学位，之后分别在清华大学、马里兰大学进行博士后研究，现为中科院计算所副研究员，硕士生导师。研究方向为计算机视觉与模式识别，重点关注复杂真实场景下的图像视频目标识别与检索等问题。目前在领域主流国际期刊和会议如IEEE TPAMI、TIP、PR、CVPR、ICCV、ICML等发表论文50余篇，Google Scholar引用1500余次。入选2012年度中科院计算所“百星计划”、2014年度微软亚洲研究院“铸星计划”、2015年度中科院青年创新促进会、2015年度“CCF-Intel青年学者提升计划”等。获得2015年度国家自然科学奖二等奖（第四完成人）。