Color Image Profiling Using Fuzzy Sets

Authors: AMINE AIT YOUNES, ISIS TRUCK, HERMAN AKDAG

Abstract: In this article, a software for image indexing and retrieval is presented. The classification proposed here is based on the dominant color(s) of the images. The process consists in assigning a colorimetric profile to the image in HLS space (Hue, Lightness, Saturation). First, the definition of hue is done thanks to a fuzzy representation to take into account the non-uniformity of colors distribution. And then, lightness and saturation are represented through linguistic qualifiers also defined in a fuzzy way. Finally, the profile is built through fuzzy functions representing the membership degree of the image to different classes. In order to improve the performances i.e. to define more accurate profiles, we propose to consider zones of pixels, instead of pixels individually. Those zones can be constructed thanks to an edge detection algorithm. A sample of pixels is chosen inside a zone to determine the color of the zone. According to the detected dominant colors, such a software may be used to classify indoor/outdoor images for example, or harmonious/disharmonious images...

Keywords: Color image classification, image profiling, dominant colors, fuzzy membership functions.

Full Text: PDF