by Mohammed Akkal, Paul Watta, Mohamad Hassoun
This program simulates an associative memory that is based on the Hamming distance similarity measure coupled with a voting scheme. Patches of an input image (starting from the center of the image) are taken and are compared to corresponding patches in a set of stored images. Matching scores of patches is kept for all images and displayed at the bottom of the applet window. The region where the patches are picked from is grown until one stored image aquires a matching score (with the input image) that is significantly higher than the rest of the stored images. The applet displays the "winning" image as the recollected image.
The user can choose the input image and can corrupt it with varying degrees of random noise.