Detecting duplicates among symbolically compressed images in a large document database

The detection of duplicate images is a useful means of indexing a large database of documents. An algorithm for duplicate document detection is proposed in this paper that operates directly on images that have been symbolically compressed using techniques related to the ongoing JBIG2 standardization e€ort. This paper describes a hidden Markov model (HMM) method that recognizes the text in an image by deciphering data from the compressed representation. Experimental results show that it can recover better than 90% of the text in compressed document images and that this is sucient to identify duplicates in a large database.