Pixel domain and compressed domain image retrieval features
Gerald Schaefer · 2013
Content-based image retrieval (CBIR) extracts features from images to support image search. In this tutorial paper, we will review some of the basic CBIR algorithms that derive colour, texture and shape features, and will then show how image features can be extracted from the compressed domain, in particular from JPEG images, without the need of completely decompressing the images.