Arts & Sciences Events
[PAST EVENT] Deep Learning for Image Manipulation
Access & Features
- Open to the public
Title: Deep Learning for Image Manipulation
Abstract: Deep neural networks are highly successful at representing complex and high dimensional data, such as images. In this talk, I will describe two examples of how such techniques can be used to capture the distribution of natural images, and based on this, to modify images according to given user objectives. In the first example, I will present a technique to address image restoration problems such as deblurring and super-resolution that builds on denoising autoencoders. Second, I will present an intuitive sketch-based image editing system that exploits conditional generative adversarial networks. My talk will illustrate how deep learning based approaches can now replace more traditional techniques in many image processing problems, while often providing superior results and opening up new opportunities for intuitive, interactive image manipulation.
Bio: Matthias Zwicker joined the University of Maryland in March 2017 as the Reginald Allan Hahne Endowed E-Nnovate Professor in Computer Science. He obtained his PhD from ETH in Zurich, Switzerland, in 2003. From 2003 to 2006 he was a post-doctoral associate with the computer graphics group at the Massachusetts Institute of Technology, and then held a position as an Assistant Professor at the University of California in San Diego from 2006 to 2008. From 2008-2017, he was a professor in Computer Science at the University of Bern, Switzerland. His research focuses on realistic rendering and modeling for computer graphics applications. He has served as a papers co-chair and conference chair of the IEEE/Eurographics Symposium on Point-Based Graphics, and as a papers co-chair for Eurographics 2010 and the Eurographics Symposium on Rendering 2017. He has been a member of program committees for various conferences including ACM SIGGRAPH, ACM SIGGRAPH Asia, and Eurographics, and he has served as an associate editor for journals such as the Computer Graphics Forum, IEEE TVCG, and the Visual Computer.
Contact
Pieter Peers