OBJECT RECOGNITION BASED ON INVARIANCE FEATURES FOR IMAGE CLASSIFICATION
Lal Kishore · 2008
Invariance is an important aspect in image object recognition. Invariance of the output with respect to certain transformations of the input is a typical example of a priori knowledge. The concept of tangent vectors, which compactly represent the essence of these transformation invariances, and two classes of algorithms “tangent distance” and “tangent propagations”, make use of these invariances to improve performance. Results obtained with an extended tangent distance incorporated in a kernel density based Bayesian classifier to compensate for affine image variations are presented in the paper. An image distortion model for local variations is introduced and its relationship to tangent distance is considered. The classification algorithms are evaluated on databases of different domains. An excellent result of 2.2% error rate on the original USPS handwritten digits recognition task is obtained. On a database of radiographs from daily routine, best results are obtained by combining tangent distance and the distortion model.