学术报告第201863期

报告题目：Fast Online Object Tracking and Segmentation: A Unifying Approach

报 告 人：张力 博士后研究员（牛津大学）

报告时间：2018年11月28日10:00

报告地点：南一楼中311

报告内容摘要：

In this talk, I will illustrate how to perform both real-time object tracking and semi-supervised video object segmentation with a single simple approach. The method, dubbed SiamMask, acts on the offline training procedure of popular fully-convolutional Siamese approaches for object tracking by augmenting their losses with a binary segmentation task. Once trained, SiamMask solely relies on a single bounding-box initialisation and operates online, producing class-agnostic object segmentation masks and rotated bounding boxes at 35 frames per second. Despite its simplicity, versatility and fast speed, our strategy allows us to establish a new state-of-the-art among real-time trackers on VOT-2018, while at the same time demonstrating competitive performance and the best speed for the semi-supervised video object segmentation task on DAVIS-2016 and DAVIS-2017.

报告人简介：

Li Zhang is Postdoctoral Research Assistant in the Department of Engineering Science at the University of Oxford, and a member of the Torr Vison Group, under the supervision of Professor Philip H.S. Torr. His research interest is on computer vision. Currently he focuses on the area of visual object tracking. He previously worked on Person Re-identification, Zero-Shot Learning, Few-Shot Learning and Reinforcement Learning for Image Captioning.