[2017-12-29] Yu-Chuan Su,University of Texas at Austin, “Learning for 360° Compression, Convolution and Presentation”

Title: Learning for 360° Compression, Convolution and Presentation 
Date: 2017-12-29  03:30 pm-04:30 pm
Location: R102, CSIE
Speaker: Yu-Chuan Su,University of Texas at Austin
Hosted by: Prof.Winston Hsu



While 360° cameras offer tremendous new possibilities in vision, graphics, and virtual reality, the spherical images they produce introduce many new challenges. In this talk, I will introduce how computer vision and machine learning can help to improve the compression, processing and presentation of the new media. For 360° video compression, we observe that some orientations of a 360° video, once projected, are more compressible than others. We introduce an approach to predict the sphere rotation that will yield the maximal compression rate. For 360° imagery processing, we propose to learn a spherical convolutional network that translates a planar CNN to process 360° imagery. Our method yields the most accurate results while saving orders of magnitude in computation versus the existing solutions. For 360° video presentation, we propose a novel algorithm for virtual cinematography in 360° video. Our algorithm produces informative videos that could conceivably have been captured by human videographers from 360° video, which free both the videographer and the end viewer from the task of determining what to watch.


Yu-Chuan Su is a Ph.D. student at the Computer Science Department at the University of Texas at Austin. His Ph.D. work studies computer vision in 360° video from various aspects, ranging from applications such as automatic view planning, learning algorithms like spherical image convolution, to the 360° data format from the image projection perspective. Before joining UT Austin in 2014, he received his M.S. from the Department of Computer Science and Information Engineering at National Taiwan University.

最後修改時間:2017-12-27 PM 4:35

cron web_use_log