文档图像基准生成系统

摘要

For the generation of base indexing information of scanned document image with noise, this system first extracts idealised indexing information based on noise-free PDF document, then registers them with the document image with noise using perspective transformation model and finally generates the base indexing information of the document image with noise. These information data are applied to test the accuracies of text recognition and retrieval. Furthermore, based on some typical different image degradation models, the system has generated the document images with different noise types in batch. Experiments show that the indexing information in this system has high accuracy, the results of image degradation are close to practical noise effect.%为生成含噪声的扫描文档图像的基准标引信息,系统首先基于无噪声的PDF文档抽取理想化标引信息,采用透视变换模型,将其与含噪声文档图像进行配准,最终生成含噪声图像的基准标引信息,将其用于测试文字识别、检索的精度.系统还基于几种经典的图像退化模型,批量产生了含不同噪声类型的文档图像.经实验表明,该系统标引信息精度高,图像退化结果与实际噪声效果接近.