Automatic detection of targets using Gabor filters and neural networks
Mircea Popa, Paul Bechet, Ştefan Demeter · 2005
Abstract: This paper presents a method for detecting artificial objects (military targets) by combining the outputs of a set of log-Gabor filters using artificial neural networks. The system is able to automatically process frequency content of an image and to extract features –visual patterns- that have a high degree of correlation in the statistical structure across frequency bands. We present some results using a filtering technique for the automatically learned partitioning of patterns in a digital image.