Fuzzy color descriptors to index Roman mosaic-images

We present efficient fuzzy color descriptors to index and retrieve images by content. The proposed approach uses fuzzy image dominant colors and fuzzy color histogram features to index images based on regions. Similar images are extracted using a fuzzy similarity and histogram intersection measures. Tests demonstrate that the fusion of the fuzzy features increase the precision of the system. The proposed index is tested on two databases: the first one containing 800 images of historical Roman mosaics from the 1st to 4th century, whereas the second database is composed of 1000 images of Corel images database. Evaluation tests and comparison with several other prevailing approaches prove a remarkable performance of our proposed approach.