Advanced search


José García-Hernández, Alejandro H. Toselli, Joaquim Arlandis, Roberto Paredes, Rafael Llobet, Alfons Juan, Juan C. Pérez-Cortés, Javier Cano, Alberto Pérez, Enrique Vidal, Francisco Casacuberta. OCR Research in PRHLT Group. Pattern Recognition: Progress, Directions and Applications, 2006. Filiberto Plá, Petia Radeva, Jordi Vitrià (Editors). pp. 106-125. Centre de Visió per Computador. ISBN 84-933652-6-2

The main purpose of this work is to provide a qualitative description of the current research area on OCR carried out by the PRHLT-ITI/DSIC research group. First, different preprocessing and features extraction methods are briefly described: tangent vectors based methods, local features extraction and others classical methods. Then, two different approaches for OCR are also presented: the k-NN method with its fast search version based on KD-Tree and the mixtures of Bernoulli classifier. Moreover, real implementations are shown at the end of this work.