Filled-in document identification using local features and a direct voting scheme
IbPRIA'11 Proceedings of the 5th Iberian conference on Pattern recognition and image analysis
| Hi-index | 0.00 |
Ordering and charging goods have been increasingly treated with faxed forms. Although the FAXOCR system for specified forms is used practically, the performance to unspecified forms is not enough, because of the effect of noise on the faxed forms during the facsimile transmission. The final target of this study is to construct a practical FAXOCR system for unspecified forms. As the first stage, an identification method for unspecified faxed forms is proposed in this paper. In our approach, character separation and position adjustment are performed in the Hough-Space as pre-processing. Then the form identification is carried out by using vote histogram in the Hough-Space. The performance of the proposed technique is verified experimentally by using actual faxed forms.