Finding Correspondences across Multiple Graphs or Images

Junchi Yan and Xiaowei Zhou  

Tutorial Abstract

This tutorial will introduce and describe the basic form of graph matching, classical two-graph matching methods, as well as some recent multi-graph matching methods. In these cases, the graph nodes are supposed to have been given in advance. Then the tutorial will extend to more practical scenarios for matching multiple images from scratch, whereby some optimization and learning methods will be covered.

Tutorial Outline

◇ Introduction to graph/image matching
◇  joint matching of multiple graphs
◇  joint matching of multiple images
◇  summary and outlook

Speaker Interlocution

Junchi Yan has been with IBM Research (Beijing, New York, Shanghai) since April 2011 and now he is the Senior Research Staff Member and IBM master inventor. He has published over 30 CCF-A papers in pattern recognition, computer vision and machine learning. He is the receiver of CCF outstanding doctorate thesis (work on graph matching). He serves as Associate Editor for IEEE ACCESS, and (Leading) Guest Editor for Pattern Recognition Letters, Multimedia Tools and Applications etc. He also serves as reviewer or PC member for JMLR/TPAMI/TIP/ TCYB/TNNLS, CVPR/ICCV/ECCV/IJCAI/AAAI/NIPS.



Xiaowei Zhou is a Research Professor in the College of Computer Science, Zhejiang University. He was a Postdoctoral Researcher in Computer and Information Science, University of Pennsylvania. His research interests are on 3D object recognition, pose estimation, shape reconstruction and correspondence problems. He cofounded the Geometry Meets Deep Learning Workshops (associated with ECCV’16 and ICCV’17) and was the organizer of the Tutorial on 3D Object Geometry from Single Image at 3DV’17. He also served as PC members or reviewers for CVPR, ICCV, ECCV, IJCAI, AAAI, TPAMI, IJCV, etc.