Efficient Object Localization for Query-by-Subregion

Localizing an object in an image is a common task in the field of computer vision, and correlograms are often used for content-based image retrieval. This paper proposes an efficient object localization algorithm for query-by-sub region. The new algorithm utilizes correlogram back-projection in the chromaticity coordinates to handle the problem of sub-region querying. This approach enables users to identify and locate primary and candidate regions without the need for further information about the object through similar spatial color information. Experimental results show that the color correlogram is markedly much more effective than a color histogram for this task.