20160622-20谢宁：Machine Learning Methods for Automatic Stroke Generation and... ...

报告嘉宾：谢宁（同济大学）

报告时间：2016年6月22日（星期三）晚20:00（北京时间）

报告题目：Machine Learning Methods for Automatic Stroke Generation and Image Stylization 

主持人：  张林（同济大学）

报告摘要： Artistic stylization in non-photorealistic rendering enables users to stylize pictures with the appearance of traditional art forms, such as pointillism painting, line sketching, or brush stroke drawing. Among them, the brush stroke drawing is one of the widely used art styles across different cultures in history. In computer-generated painterly rendering, the stroke placement is a big challenge and significant efforts have been made to investigate how to draw a stroke with realistic brush texture in a desired shape and how to organize multiple strokes. 

   We believe that the human behavior for artistic creativity contents more information on the artistic stylization. In this talk, we will present several techniques that significantly contribute to dynamic stroke-based rendering (SBR) towards machine learning methods, especially how to model the brush agent under reinforcement learning framework. We implement a web application for agent's automatic stylization of converting photographs into stroke drawings. We evaluate availability and effectiveness of our methods with the mainstream commercial softwares (Adobe Photoshop, Corel Painter, and Celsys Retas Studio) through the user study. Rendering results show that our methods can successfully draw complex shapes with smooth and natural brush strokes. In addition, applications to automatic photo conversion into an Oriental ink painting are demonstrated to be promising.

报告人简介：谢宁，工学博士，吉大本科，东京工业大学硕士、博士（导师：Masashi Sugiyama URL:http://www.ms.k.u-tokyo.ac.jp/）。2014年10月入职同济大学软件学院数字媒体系，任职：助理教授。近年来，专注于非真实感渲染（NPR），艺术风格化绘制（Artistic rendering）和增强学习（Reinforcement Learning）等研究。目前，在ICML、IJCAI、Siggraph Asia、NPAR等顶级会议发表论文多篇。(主页：http://sse.tongji.edu.cn/xiening/)