"IMAGE DIGEST III: A NEAREST NEIGHBOUR DIFFERENTIAL BASED IMAGE DIGEST GENERATION ALGORITHM "
Sesha Kalyur, G. S. Nagaraja · IJITR International Journal of Innovative Technology and Research - IJITR International Journal of Innovative Technology and Research · 2015
In this paper we present a methodology to generate a digest for an image based on the grayness value differentials that exist between neighboring pixels. Neighboring pixels are those that lie to the immediate left, immediate right, immediate above and immediate below of a given pixel plus the four pixels that lie in between. This algorithm works on the monochrome images of VGA resolution. Color images are converted to their monochrome equivalents. Images of resolution higher than VGA are converted to images of VGA resolution. The given image is divided into equal sized segments or regions. The pixels of the given image are sampled in such a way that each segment contributes equally to the sampled set for the image. This algorithm uses a histogram based statistical approach towards digest generation. Counters are maintained at the segment level, which keep the raw counts of the differentials for the sampled pixels. The counter values are composed to form the digest for the segment. Computing the digest at the segment level helps to preserve the locality information for the image. The digest for the entire image is a composition of the individual digests generated for each segment or region. The method also provides for the calculation of a lite version of the digest that saves digest space by ignoring the region or locality information.