Independent component analysis applied to feature extraction from colour and stereo images

Previous work has shown that independent component analysis (ICA) applied t o feature extraction from natural image data yields features resembling Gab or functions and simple-cell receptive fields. This article considers the e ffects of including chromatic and stereo information. The inclusion of colo ur leads to features divided into separate red/green, blue/yellow, and brig ht/dark channels. Stereo image data, on the other hand, leads to binocular receptive fields which are tuned to various disparities. The similarities b etween these results and the observed properties of simple cells in the pri mary visual cortex are further evidence for the hypothesis that visual cort ical neurons perform some type of redundancy reduction, which was one of th e original motivations for ICA in the first place. In addition, ICA provide s a principled method for feature extraction from colour and stereo images; such features could be used in image processing operations such as denoisi ng and compression, as well as in pattern recognition.