Semantic labeling of images combining color, texture and keywords

Content-Based Image/Video Retrieval Systems combine perceptual features such as color, texture and shape with semantic concepts for improving the quality of the query's results. In this paper, a retrieval technique combining color and texture with keywords is presented. A colorbased method and conjunction with a mining association rules technique are used to propagate labels from a subset of annotated images into a large-scale images database. A texture-based method is applied to associate semantic labels (keywords) to regions within the image. This technique is useful to retrieve sub-images from an images database using text-based queries.