20160817-27沈小勇：Automatic Portrait Segmentation and Matting

报告嘉宾1：沈小勇（香港中文大学）

报告时间：2016年8月17日（星期三）晚20:00（北京时间）

报告题目：Automatic Portrait Segmentation and Matting

主持人：王乃岩（TuSimple）

报告摘要：In this talk, I will introduce our fully automatic segmentation and matting methods for portrait images. The method does not need any user interaction, which was however essential in most previous approaches. In order to accomplish this goal, a new end-to-end convolutional neural network (CNN) is proposed taking the input of a portrait image. It outputs the segmentation or mating results. Our CNN considers not only image semantic prediction but also pixel-level matte optimization. In addition, a new portrait image dataset is constructed with our labeled segmentation and matting ground truth. Our automatic method achieves comparable segmentation and matting results with state-of-the-art methods that require specified foreground and background regions or pixels. Many applications are enabled given the automatic nature of our system.

参考文献：

[1] Xiaoyong Shen, Aaron Hertzmann, Jiaya Jia, Sylvain Paris, Brian Price, Eli Shechtman, Ian Sachs, "Automatic Portrait Segmentation for Image Stylization"Computer Graphics Forum, 35(2), 2016 ( Proc. Eureographics 2016) .

[2] Xiaoyong Shen, Xin Tao, Hongyun Gao, Chao Zhou, Jiaya Jia, "Deep Automatic Portrait Matting" European Conference on Computer Vision (ECCV) , 2016. (Spotlight Presentation)

报告人简介：

Xiaoyong Shen received the PhD degree in Computer Science and Engineering Department in the Chinese University of Hong Kong. His supervisor is Prof. Jiaya Jia. Before that, He received the B. S. degree in Computational Mathematics and M. S. degree in Applied Mathematics from Zhejiang University in 2010 and 2012 respectively, under the supervision of Prof. Ligang Liu. His research interest includes computer vision and computer graphics, especially on image filtering, image restoration, image matching, object detection and segmentation.

[Slides]