20170524-10 许佳：DC Flow

报告嘉宾：Jia Xu（Intel Visual Computing Lab）

报告时间：2017年05月24日（星期三）上午10:00（北京时间）

报告题目：DC Flow

主持人：禹之鼎（Carnegie Mellon University）

报告摘要：

In this talk, I will present a new optical flow estimation approach that operates on the full four-dimensional cost volume. This direct approach shares the structural benefits of leading stereo matching pipelines, which are known to yield high accuracy. To this day, such approaches have been considered impractical due to the size of the cost volume. We show that the full four-dimensional cost volume can be constructed in a fraction of a second due to its regularity. We then exploit this regularity further by adapting semi-global matching to the four-dimensional setting. This yields a pipeline that achieves significantly higher accuracy than state-of-the-art optical flow methods while being faster than most. Our approach outperforms all published general-purpose optical flow methods on the KITTI 2015 benchmark, and has been #1 on the Sintel benchmark since 10/2016.

参考文献：

Jia Xu, René Ranftl, and Vladlen Koltun. Accurate Optical Flow via Direct Cost Volume Processing. In Computer Vision and Pattern Recognition (CVPR), July 2017.

报告人简介：

Jia Xu is a Senior Research Scientist at the Intel Visual Computing Lab. His main research interests include computer vision and deep learning, with a particular focus on visual perception (e.g., semantic segmentation, dense visual correspondence, video analytics/segmentation). Jia received his Ph.D. in Computer Sciences from the University of Wisconsin-Madison in 2015. He was also a visiting student in TTI-Chicago and University of Toronto during 2013 and 2014. Prior to graduate school, Jia obtained his B.S. from Nanjing University in 2010.