【摘要】 由于汉字笔画复杂,从视频中提取的汉字图像质量往往较差,采用传统光学字符识别(OCR)的结果不理想.为了解决低质量汉字图像的识别问题,提出一种基于分块搜索的两级识别方法.首先建立汉字图像的分块结构并模仿低质量汉字生成训练集,然后对训练集中各分块图像应用主成分分析提取特征并建立索引.待识别图像应用分块搜索和投票的方式从索引中获取候选汉字集合(一级识别),再根据投票结果的显著性辅以全局结构特征匹配识别汉字(二级识别).实验结果证明,该方法对于低质量汉字图像比普通的OCR方法具有更高的识别率.更多还原

【Abstract】 Due to the complex character strokes,the quality of video-extracted Chinese character images is often poor,for which traditional optical character recognition(OCR) could not get desired results.To address this problem,this paper presents a two-stage scheme for low-quality Chinese character recognition based on block search.The block structure of the Chinese character image is built,along with a training set,imitating low-quality Chinese characters.And an index is generated after extracting the features from each block of the training set by applying principle component analysis.This scheme retrieves candidate character set from the index by block search and voting(1st stage),and then recognizes the character according to the salience of voting result assisted with global structural feature match(2nd stage).The experimental results have demonstrated that this scheme has better recognition rate for low-quality Chinese character images compared with traditional OCR method.更多还原