Section: Image Processing and Pattern Recognition

Combined approach to localization of differences for multimodal images

D. M. Murashov

Dorodnitsyn Computing Centre of the Russian Academy of Sciences

Abstract: An approach to the problem of localization of the differences in multimodal images is suggested. The approach is based on specific object detectors and local information-theoretical image difference measures implemented as conditional entropy of the analyzed image pair. The proposed approach is applied to the problem of detecting repainting areas of fine art paintings using the images acquired in visible and ultraviolet spectral ranges.

Keywords: multimodal images; measure of image difference; information-theoretical measure; conditional entropy; images of fine art paintings.