The Use of Grouping in Visual Object Recognition
Abstract
This report explores the use of grouping in object recognition by computational systems. Many recognition systems in the past have performed an undirected search through the space of different segmentations of an image in order to recognize objects. This approach leads to significant problems of computational complexity and accuracy. The process of grouping determines the sections of an image most likely to come from a single object. This can tell a recognition system which segmentations of the image to consider first, improving both its speed and accuracy. The report describes a particular recognition system called GROPER. GROPER performs grouping by using distance and relative orientation constraints that estimate the likelihood of different edges in an image coming from the same object. The thesis presents both a theoretical analysis of the grouping problem and a practical implementation of a grouping system. And it discusses the relevance of this theory of grouping to human psychology. GROPER also uses an indexing module to allow it to make use of knowledge of different objects, any of which might appear in an image. We test GROPER by comparing it to similar recognition system that does not use grouping.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1988
- Accession Number
- ADA201691
Entities
People
- David W. Jacobs
Organizations
- Massachusetts Institute of Technology