In this paper, we propose a new approach of document analysis and recognition (DAR) based on the combination of OCR systems. The proposed approach aims to improve the document recognition by combining the result of several OCR systems according to their performances. We focus our attention, in this paper, on the table combination. We start by presenting the results of the evaluation of OCR system in the table block recognition. Then, we present our table combination method. We are interested in both table structure and table content.
Citation:
Ferihane Kboubi, Anja Habacha Chabi, Mohamed Ben Ahmed, "Table Recognition Evaluation and Combination Methods," icdar, pp.1237-1241, Eighth International Conference on Document Analysis and Recognition (ICDAR'05), 2005