20160727-24 Zihan Zhou:Towards Automatic Composition Modeling...

报告嘉宾1：Zihan Zhou（Penn State University, USA）

报告时间：2016年7月27日（星期三）晚20:00（北京时间）

报告题目：Towards Automatic Composition Modeling and Feedback for Photographers

主持人：韩军伟（西北工业大学）

报告摘要：

The capacity of automatically modeling photographic composition is valuable for many real-world applications such as digital photography, image retrieval, image understanding, and image aesthetics assessment. In this talk, I will discuss our recent work on composition modeling that aims to provide on-site guidance to amateur photographers on their work. Specifically, we propose a new method to analyze the use of linear perspective in the photo by partitioning the photo into photometrically and geometrically consistent regions. Comparing with existing methods that focus on basic rules of design such as visual balance, golden ratio, and the rule of thirds, our new quantitative model is more comprehensive when perspective is relevant. Based on the model, we further develop a novel image retrieval system that is able to find exemplar photos with similar viewpoints as the query photo from a collection of photos taken by experienced photographers. We show that these exemplars could serve as an informative guide for the users to achieve good compositions in their photos.

参考文献：

[1] Zihan Zhou, Siqiong He, Jia Li, and James Z. Wang. Modeling Perspective Effects in Photographic Composition. ACM Multimedia Conference (ACM MM), 2015.

[2] Zihan Zhou*, Siqiong He*, Jia Li, and James Z. Wang. Modeling Photographic Composition via Triangles. arXiv:1605.09559, 2016. (*=equal contributions)

报告人简介：

Dr. Zihan Zhou is currently a faculty member in the College of Information Sciences and Technology at the Pennsylvania State University. Before joining Penn State, he completed his Ph.D in Electrical and Computer Engineering at the University of Illinois at Urbana-Champaign in 2013. In Summer 2016, he is a visiting professor in the School of Information Science and Technology, ShanghaiTech University. His research interest lies in the broad area of computer vision, image processing, and machine learning, with a focus in developing and analyzing new computational tools for efficient and robust discovery of low-dimensional data structure from high-dimensional observations. Recently, he is particularly interested in exploiting scene structure regularities for 3D reconstruction and its applications. To learn more, please visit his homepage: https://faculty.ist.psu.edu/zzhou