A Spatial Frequency Analysis Model for Predicting Human Performance at Visual Pattern Matching Tasks,
Abstract
A model for simulating human performance in visual pattern matching tasks is presented. The model is based on evidence of spatial frequency processing in the visual system, and on the hypothesis that shape recognition is determined only by the low spatial frequency harmonics of the image. Two psychophysical pattern matching experiments are described that demonstrate a clear functional relationship between the 'similarity' of two patterns as judged by human observers and the Euclidean distance between spatially filtered Fourier transforms of the patterns. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1977
- Accession Number
- ADA041649
Entities
People
- Mark W. Cannon Jr.
Organizations
- Air Force Research Laboratory