Noise tolerant object recognition using Gabor filtering

The choice of features for invariant object recognition is one of the most essential problems in computer vision. The authors have previously proposed Gabor (1946) filtering based feature extraction methods which have been successfully applied in invariant object recognition. In this study, the Gabor filtering based feature extraction is further analysed in terms of distortion tolerance which is an essential property for many applications. Experiments indicate that an accurate recognition can be achieved in the presence of significant amounts of distortions.