Course info

Authors: Remco Veltkamp.

Level: Advanced.

Prerequisites: Basic computer graphics and image processing knowledge.

Description: Shape matching is a central problem in visual information systems, computer vision, pattern recognition, and robotics. Applications of shape matching include industrial inspection, fingerprint matching, and content-based image retrieval. A few examples are:

• Applications in agricultural inspection. A typical problem here is to find a matching transformation. Based on shape characteristics, we can find the transformation that matches one piece of fruit with another.
• Point matching in fingerprint identification applications. After extraction of featuring points, two point sets must be matched. The difficulty here is that there is typically no one to one correspondence between the two point sets. The matching technique should be robust against noise and occlusion.
• Application in multimedia retrieval. Given the query shape, the task is to find all pictures that contain similar shapes. The typical problem is that only pieces of the query shape appear in only parts of some of the database pictures.

This tutorial is about the matching of shape, both in 2D and in 3D. We will survey:

• Classes of applications
• Various matching problems
• Shape features
• Similarity measures
• Algorithms

Literature:
Remco C. Veltkamp, Michiel Hagedoorn. State-of-the-art in shape matching. In: Principles of Visual Information Retrieval, M. Lew (ed.), Springer, 2001, ISBN 1-85233-381-2, 87-119.
[pdf file]

Format: Slides in pdf-format.

Course material

1. Introduction
2. Perception
3. Feature-based Matching
4. Direct Matching
5. 3D Shape Matching