%0 Journal Article
%T Multimodal page classification in administrative document image streams
%A Marçal Rusiñol
%A Volkmar Frinken
%A Dimosthenis Karatzas
%A Andrew Bagdanov
%A Josep Llados
%J International Journal on Document Analysis and Recognition
%D 2014
%V 17
%N 4
%I Springer Berlin Heidelberg
%@ 1433-2833
%F Marçal Rusiñol2014
%O DAG; LAMP; 600.056; 600.061; 601.240; 601.223; 600.077; 600.079
%O exported from refbase (http://refbase.cvc.uab.es/show.php?record=2523), last updated on Wed, 04 Feb 2015 16:52:34 +0100
%X In this paper, we present a page classification application in a banking workflow. The proposed architecture represents administrative document images by merging visual and textual descriptions. The visual description is based on a hierarchical representation of the pixel intensity distribution. The textual description uses latent semantic analysis to represent document content as a mixture of topics. Several off-the-shelf classifiers and different strategies for combining visual and textual cues have been evaluated. A final step uses an n-gram model of the page stream allowing a finer-grained classification of pages. The proposed method has been tested in a real large-scale environment and we report results on a dataset of 70,000 pages.
%K Digital mail room
%K Multimodal page classification
%K Visual and textual document description
%U http://dx.doi.org/10.1007/s10032-014-0225-8
%P 331-341