Gradient descent techniques for feature detection template generation

William F. Pont Jr.; Paul D. Gader

doi:10.1117/12.46120

Translator Disclaimer

1 July 1991 Gradient descent techniques for feature detection template generation

William F. Pont Jr., Paul D. Gader

Author Affiliations +

Proceedings Volume 1568, Image Algebra and Morphological Image Processing II; (1991) https://doi.org/10.1117/12.46120
Event: San Diego, '91, 1991, San Diego, CA, United States

Abstract

An algorithm is described for estimating the weights of gray-scale templates used for morphological feature detection. The problem of finding optimal templates for object detection is formulated as a statistical estimation problem. The problem is solved using a gradient descent algorithm. This procedure is applied to the problem of detecting trucks in simulated SAR imagery. Good performance is achieved with 22 of 24 trucks being identified in a test set and no false alarms.

Citation Download Citation

William F. Pont Jr. and Paul D. Gader "Gradient descent techniques for feature detection template generation", Proc. SPIE 1568, Image Algebra and Morphological Image Processing II, (1 July 1991); https://doi.org/10.1117/12.46120

ACCESS THE FULL ARTICLE