Effective Image Fingerprint Extraction Based on Random Bubble Sampling

In this paper we propose an algorithm for image fingerprint extraction based on random selection of circular bubbles on the considered image. In more detail, a fingerprint vector is associated to the image, the components of which are the variances of pixel luminance values in randomly selected circular zones of the image. The positions and radius of these bubbles result from a random selection, whose parameters are user-defined. The obtained fingerprint has then been used for content-based image retrieval, using the standard euclidean distance as similarity metric between the extracted features. Experiments based on the detection of various linearly and nonlinearly distorted versions of a test image in a large database show very promising results.