Object Recogntition with Representations Based on Sparsified Gabor Wavelets Used as Local Line Detectors

We introduce an object recognition system (called ORAS-SYLL) in which objects are represented as a sparse and spatially organized set of local (bent) line segments. The line segments correspond to binarized Gabor wavelets or banana wavelets, which are bent and stretched Gabor wavelets. These features can be metrically organized, the metric enables an eecient learning of object representations. Learning can be performed autonomously by utilizing motor{controlled feedback. The learned representation are used for fast and eecient localization and discrimination of objects in complex scenes. ORASSYLL has been heavily innuenced by an older and well known vision system 4, 9], and has also been innuenced by Biederman's comments to this older system 1]. A comparison of ORASSYLL and the older system, including some remarks about the speciic role of Gabor wavelets within ORASSYLL, is given at the end of the paper.