GAMES Webinar 2019 – 101期 （几何专题课程）|Yukun Lai（Cardiff University），傅孝明（中国科学技术大学）
【GAMES Webinar 2019-101期】
报告嘉宾：Yukun Lai，Cardiff University
报告题目：Non-photorealistic Rendering of Images: from Image Processing to Machine Learning
Given input inages, non-photorealistic rendering (NPR) aims to produce stylized images that often mimic specific artistic effects such as line drawings, oil painint, watercolor, etc. It is a topic that has been studied for decades. Traditional NPR research develops dedicated image processing algorithms for specific styles. In recent years, non-photorealistic rendering using machine learning, or so-called neural style transfer, is becoming popular. This talk will overview several NPR techniques we developed, with a focus of producing minimalistic styles that are highly abstract, which tend to be challenging for algorithms. I will discuss and compare techniques based on both traditional image processing, and recent development based on deep learning.
Dr Yu-Kun Lai is a Reader in the Visual Computing group, School of Computer Science and Informatics, Cardiff University, UK. He obtained his PhD degree from Tsinghua University, China in 2008 and joined Cardiff University shortly after that. He has been working on broad areas of Visual Computing, including Computer Graphics, Geometric Processing, Image Processing and Computer Vision. He has published over 100 papers in world-class journals and conferences, including ACM TOG, IEEE TVCG, IEEE TIP and CVPR etc. He is on the editorial boards of Computer Graphics Forum and The Visual Computer, conference co-chair of SGP 2014 and CVM 2016, and on the programme committee of many major international conferences.
报告题目：Atlas generation: chart generation, chart parameterizations, and chart packing
Atlas generation consists of three main components: chart generation, chart parameterizations, and parameterized chart packing. In this talk, I will introduce our works on these three problems: (1) a sphere-based algorithm for computing high-quality cuts for generating low isometric distortion planar parameterizations; (2) a novel approach, called Progressive Parameterizations, for computing foldover-free parameterizations with low isometric distortion on disk topology meshes; (3) a novel algorithm for refining an input atlas with bounded packing efficiency.
GAMES主页的“使用教程”中有 “如何观看GAMES Webinar直播？”及“如何加入GAMES微信群？”的信息；