20170427-08 Yong Jae Lee：Discovering where to look for...

报告嘉宾：Yong Jae Lee（UC Davis Computer Science）

报告时间：2017年04月27日（星期四）上午10:30（北京时间）

报告题目：Discovering where to look for weakly-supervised visual recognition

主持人：禹之鼎

报告摘要：I will first give a brief introduction of the computer vision group at UC Davis, including the research directions of each member. I will then present our group's recent and ongoing work on weakly-supervised visual recognition. In contrast to fully-supervised algorithms, the proposed methods do not require detailed annotations during training, and instead can learn to focus on the relevant visual regions given only image-level or video-level semantic tags that state whether an object is present or absent (e.g. an image tagged with "car"). I will show how the proposed algorithms can produce state-of-the-art weakly-supervised results for object localization and detection.

参考文献：

[1] Hide-and-Seek: Forcing a Network to be Meticulous for Weakly-supervised Object and Action Localization Krishna Kumar Singh and Yong Jae Lee In submission, Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017 

[2] Multi-task Feature Learning using Synthetic Game Imagery Zhongzheng Ren and Yong Jae Lee In submission, Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017 

[3] Weakly-supervised Visual Grounding of Phrases with Linguistic Structures Fanyi Xiao, Leonid Sigal, and Yong Jae Lee Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017 

[4] Interspecies Knowledge Transfer for Facial Keypoint Detection Maheen Rashid, Xiuye Gu, and Yong Jae Lee Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017 

[5] Track and Transfer: Watching Videos to Simulate Strong Human Supervision for Weakly-Supervised Object Detection Krishna Kumar Singh, Fanyi Xiao, and Yong Jae Lee Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016 

[6] Track and Segment: An Iterative Unsupervised Approach for Video Object Proposals Fanyi Xiao and Yong Jae Lee Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016

报告人简介：

Yong Jae Lee is an Assistant Professor in the Department of Computer Science at the University of California, Davis. His research in computer vision and machine learning focuses on object recognition and video understanding. Before joining UC Davis in 2014, he spent one year as a post-doc at UC Berkeley and one year as a post-doc at Carnegie Mellon University, both under the supervision of Alyosha Efros. He received his PhD from the University of Texas at Austin in 2012 under the supervision of Kristen Grauman and his BS from University of Illinois at Urbana-Champaign in 2006.