Image Representation, Matching, and Recognition Using Invariant Region Descriptors

Speaker :Professor Zen Chen,
National Chiao-Tung UniversityAbstract: In 3D computer vision a scene or object in the real world can be characterized by multiple views imaged under different viewpoints and illuminations. The spatial and temporal relationships across these views are important to scene analysis and understanding. The local features are generally more robust to viewpoint change than the global features. The region feature design must take these into consideration, too. In the talk, topics related to local image representation, matching, and recognition under the geometric and photometric variations are addressed. First, a number of popular region descriptors will be reviewed and a new image descriptor, called the Zernike moment (ZM) phase, is to be introduced. The discriminative power of the new ZM phase descriptor is compared with five major existing region descriptors based on the precision-recall criterion. The experimental results involving more than 15 million region pairs indicate the proposed ZM phase descriptor has, overall speaking, the best performance under the common photometric and geometric transformations. Second, the proposed ZM phase descriptor is extended to present a new recognition method of logos imaged by the mobile phone cameras. The logo recognition can be incorporated with mobile phone service tools for use in enterprise identification, corporate website access, traffic sign reading, security check, and the related applications. Third, the one-to-one correspondence matching of feature points and its application to view registration will be addressed. We take advantage of preprocessing of the reference image offline to gather the important statistics for guiding view registration. Finally, an online registration process is presented to estimate a transformation matrix to overlay the reference image over a sensed imageBio: Prof. Zen Chen is Professor of Computer Science at National Chiao Tung University, Hsin-Chu, Taiwan. He holds a PhD from Purdue University in Electrical Engineering. Besides teaching at National Chiao Tung University, Prof. Chen spent one year (1981-1982) at Lawrence Berkeley Laboratory, University of California, Berkeley, California, as a visiting scientist and a half year (August 1989 to January 1990) at the Center for Automation Research, University of Maryland, College Park, Maryland, as a visiting professor. Prof. Chen performs research in computer vision, pattern recognition, virtual reality, and parallel algorithms and architectures. Prof. Chen is a Fellow of International Association for Pattern Recognition (IAPR). He was the founding President of the Chinese Society of Image Processing and Pattern Recognition in Taiwan, a member society of IAPR. He received the Outstanding Engineering Professor Award from the Chinese Institute of Engineers and the Outstanding Research Awards from the National Science Council of Taiwan. He is also a consultant to a number of industrial research institutes and companies in Taiwan.

Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248

Audiences: Everyone Is Invited

Contact: Talyia Veal