Comparative Study of Online Color-Based Image Retrieval Algorithms in Medical Applications
Florin Perianu, Vasile Cornita, Rodica Strungaru, Sever Pasca and Mihaela Ungureanu
The article is a comparative study of content based image retrieval by color methods. In order to compare those methods of content based image retrieval, a software system that allows the querying and then on-line actualization of a database with color images was designed. The system allows two ways of database querying: simple query and comparative query, both taking into consideration the color. The images are preprocessed, namely they are transformed from the RGB color space to HSV and CIE Luv color spaces and quantized, being thus prepared for a future query. Also, the possibility to use neural network algorithms in order to auto-select the best suited retrieval algorithm is explored.
Keywords: Content based image retrieval algorithms, quantization, color spaces, image query, similitude between images.