基于模糊推理的名片识别后处理方法

Post-process method for business card recognition based on fuzzy reasoning

针对名片信息的特点和通常的OCR名片识别方法识别率较低的问题,提出一种新的基于模糊推理的名片识别后处理方法.该方法通过OCR识别得到的文本信息和候选文本信息来进行文本内容分析,通过OCR过程中的图像切分参数进行版面分析,在分析中均采用模糊推理的方法,同时,提出一种新的模糊运算的交型算子,应用于模糊推理运算中.最后综合上述内容分析和版面分析的结果得到最终的信息分类结果.实验结果表明,该方法在名片识别和分类正确率方面明显优于其他几种常用名片系统采用的算法,本方法不仅提高了OCR识别的正确率,而且还提高了经后处理以后的识别正确率.

in view of the characteristics of card information, a new post - process method for business card recognition based on fuzzy reasoning is proposed to solve the low rate of card recognition with the OCR method. The card information is analyzed according to OCR recognition results and candidate results. The typeset page is analyzed by image-syncopated parameters provided by the OCR system. In the above process, the fuzzy reasoning method is applied to the typeset analysis. Furthermore, a new type of intersection operator is designed which is used in fuzzy reasoning. Then the final results are obtained by fusing the above information. The experimental results demonstrate that this method is effective in improving the accuracy rate of OCR recognition and classification of business cards, including the post-process results when compared with other methods.