Introduction
Scope and target audience
Visual correspondence is a key problem in many computer vision and pattern recognition tasks. The past decades have witnessed the rapid expansion of the frontier for automatic correspondence establishment among images/graphics, which is largely due to the advances in computational capacity, data availability and new algorithmic paradigms. Although the visual correspondence problem has been extensively studied in the context of multi-view geometry, its more generalized forms, along with underlying connections among different methods and settings, have not been fully explored. Meanwhile, the combination of big visual data and the deep learning paradigm has achieved significant success in many perceptual tasks; however, the existing paradigm is still far from a panacea to the correspondence problem, which often calls for more careful treatments on the local and global structures. In this workshop, we attempt to assemble recent advances in the correspondence problem, considering the explosions of big visual data applications and the deep learning algorithms.
Tracks

Graph matching and image registration

  1. Graph representation and modeling by using image/ graphics data
  2. Robust matching/registration theory and approaches for establishing visual correspondences over two or more images/graphics
  3. Partial, one-to-many or many-to-many matching models and algorithms, especially with major noise and outliers
  4. Similarity between graphs/graphics and graph clustering/classification
  5. Cross-network matching

Tracking and optical flow

  1. Multiple object tracking and association
  2. Robust and/or efficient optical flow methods
  3. Visual trajectory analytics
  4. Person Re-ID

Correspondence for 3-D vision

  1. Calibration, pose estimation and visual SLAM
  2. Depth estimation and 3-D reconstruction

Learning for/by matching

  1. Learning graph structure and similarity from data with established or unestablished correspondences
  2. Learning image feature representation from established or loosely established correspondence
  3. Common/similar objects discovery and recognition from images

Applications

Application of correspondence technology to solve any real-world image understanding problems including object detection/recognition among images/graphics, image stitching, 2-D/3-D recovery, robot vision, photogrammetry and remote sensing, industrial imaging, embed system etc.
Invited Speakers
Image based camera localization towards challenging problems
Yihong Wu
Chinese Academy of Science
Finding consistent feature correspondences across multiple images
Xiaowei Zhou
Zhejiang University
Tacking based text detection and recognition from scene and web videos
Xu-Cheng Yin
University of Science and Technology Beijing
Main Organizers
  • Junchi Yan
    Shanghai Jiao Tong University
    yanjunchi@sjtu.edu.cn
  • Shuhan Shen
    Institute of Automation, CAS, China
    shshen@nlpr.ia.ac.cn
  • Changsheng Li
    UESTC, China
    lichangsheng@uestc.edu.cn
  • Yan Zhu
    Facebook, USA
    yzhu@fb.com
  • Yinqiang Zheng
    NII, Japan
    yqzheng@nii.ac.jp
  • Xiaoyong Pan
    Erasmus MC, Netherlands
    x.pan@erasmusmc.nl
Committees
Details
It is a half-day workshop.And paper will not be published as proceedings. Posters are welcomed for exchange onsite.
Schedule
9:05 – 9:10
Opening remarks
9:10 – 9:40
Invited Speaker 1 – Yihong Wu – Image based camera localization towards challenging problems
9:45 – 10:15
Invited Speaker 2 – Xiaowei Zhou - Finding consistent feature correspondences across multiple images
10:20 – 10:50
Invited Speaker 3 – Xu-Cheng Yin - Tacking based text detection and recognition from scene and web videos
10:50 – 12:00
Coffee break + posters