Translator Disclaimer
6 June 1995 Computational vision models of early vision for target acquisition
Author Affiliations +
Proceedings Volume 2426, 9th Meeting on Optical Engineering in Israel; (1995)
Event: Optical Engineering in Israel: 9th Meeting, 1994, Tel-Aviv, Israel
Target acquisition methodology for infrared (IR) and visual man-in-the-loop imaging sensors has several limitations for many sensor performance assessment applications. Recent advances in computational vision modeling (CVM) have made dramatic improvements in the understanding of early human vision processes. A simple model of neural receptive fields consists of a generic image representation of the spatial processing characteristics for early vision cortical areas. The input image is first divided into its three color opponent components with each axis further decomposed into a set of band pass spatial frequency filters with different center frequencies and orientations. The spatial frequency decomposition is accomplished by an efficient encoding algorithm incorporating a hierarchical cascading Gaussian pyramid algorithm which is an alternating sequence of image output passing through Nyquist low pass spatial filter and subsampling local operators for image encoding. This paper examines the limitations of earlier target acquisition models and describes a computational model which starts with actual stimulus images as an input. It predicts human performance of experimental tasks by attaching a signal-to-noise ratio (SNR) to each spatial frequency channel, and then uses a combining function to define a composite d' parameter for a signal detection theory calculation of probabilities of detection and false alarm. Several examples of the model are applied to various detection applications.
© (1995) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Grant R. Gerhart, Thomas J. Meitzler, and Eui Jung Sohn "Computational vision models of early vision for target acquisition", Proc. SPIE 2426, 9th Meeting on Optical Engineering in Israel, (6 June 1995);


Psychophysical Rating Of Image Compression Techniques
Proceedings of SPIE (August 15 1989)
Early vision model for target detection
Proceedings of SPIE (May 22 1995)
Network compensation for missing sensors
Proceedings of SPIE (June 01 1991)

Back to Top