Emulating human visual perception for measuring difference inimages using an SPN graph approach

摘要

This paper presents a new methodology for efficiently representingnthe content of images and comparing images by detecting and recordingntheir visual differences. In particular, the methodology presented herenis based on a stochastic Petri-net (SPN) graph approach able to extractnand record local and global features from both images, compare them, andndefine the percentage of similarity. One of the features of the humannvisual perception is the detection of similarities between two images.nThe visual similarity is based on color, size, shape, and local andnglobal topological changes of the image regions. Several methods dealingnwith image or object similarities have been proposed. The new feature ofnthe method here is the partial emulation of the human observer's visualnperception by recording differences extracted from different images.nResults of the method described here are presented for a variety ofnimages by using local and global noisy conditions