Journal of Artificial Intelligence Research 3 (1995)187-222
ReSubmitted 5/95; published 10/95
(c) 1995 AI Access Foundation and Morgan Kaufmann Publishers. All rights reserved.

Learning Membership Functions in a Function-Based Object Recognition System

Kevin Woods woods@bigpine.csee.usf.edu
Computer Science & Engineering
University of South Florida
Tampa, FL 33620-5399

Diane Cook cook@centauri.uta.edu
Computer Science & Engineering
University of Texas at Arlington
Arlington, TX 76019

Lawrence Hall hall@waterfall.csee.usf.edu
Kevin Bowyer kwb@bigpine.csee.usf.edu
Computer Science & Engineering
University of South Florida
Tampa, FL 33620-5399

Louise Stark stark@napa.eng.uop.edu
Electrical and Computer Engineering
University of the Pacific
Stockton, CA 95211

Abstract:

• Introduction
• Related Work
• The GRUFF Object Recognition System
  • The Knowledge Primitives
  • The Category Definition Tree
  • Combination of Evidence
• The OMLET Learning System
  • Error Propagation
  • Initial Estimate of Measurement Functions
  • Adjusting Membership Functions
  • The Training Approach
    • One Disjunct Per Lesson
    • Practical Justification
• Experimental Setup
  • Test on the GRUFF Chair Database
  • Test on a Synthetic Cup Database
  • Learning from Human Evaluation Measures
• Experimental Results
  • The GRUFF Chair Database
  • The Synthetic Cups Database
  • The Chair Database for Human Evaluation
• Summary and Discussion
• References
• About this document ...