This paper presents an image retrieval framework that uses affine image moment invariants as descriptors of local image areas. Detailed feature vectors are generated by feeding the produced moments into a Bag-of-Visual-Words ...