Binary Factorization

Computing similarity measure between images is a basic task required for image classification and matching. The technique proposed explores the Moire phenomenon for determining the similarity measure for two images. When two similar images are superimposed the concentric Moire rings point to the location of the fixed point. A Hough based technique is used for circle detection. The algorithm is successfully applied to the face recognition problem.

The unsupervised learning of feature extraction in high dimensional patterns is a central problem for neural network approach. Feature extraction is the procedure which maps original patterns into the features (or factors) space of reduced dimension. In this paper we demonstrate that Hebbian learning in Hopfield-like neural network is a natural procedure for unsupervised learning of feature extraction. Due to this learning, factors become the attractors of network dynamics, hence they can be revealed by the random search. The neurodynamics is analyzed by Single-Step approximation, which is known [1] to be rather accurate for sparsely encoded Hopfield network. Thus, the analysis is restricted by the case of sparsely encoded factors. The accuracy of Single-Step approximation is confirmed by computer simulations.

When the programmable digital computer was born shortly before mid-century, there was little reason to expect that it would someday be used to write letters, keep track of supermarket inventories, run financial networks, make medical diagnoses, help design automobiles, play games, deliver e-mail and photographs across the Internet, orchestrate battles, guide humans to the moon, create special effects for movies, or teach a novice to type. In the dawn years its sole purpose was to reduce mathematical drudgery, and its value for even that role was less than compelling. One of the first of the breed was the Harvard Mark I, conceived in the late 1930s by Harvard mathematician Howard Aiken and built by IBM during World War II to solve difficult ballistics problems. The Mark I was 51 feet long and 8 feet high, had 750,000 parts and 500 miles of wiring, and was fed data in the form of punched cardsÐ'--an input method used for tabulating equipment since the