Local features-based script recognition from printed bilingual document images
International Journal of Computer Applications in Technology
Hi-index | 0.00 |
Script identification prior to OCR is necessary in document image analysis. And each script has unique spatial distribution and visual attribute that make it possible to identify itself from other languages. The key technology of script identification algorithm is to abstract effective measure feature. By analyzing vision differences based on normalized histogram statistic, Chinese, Japanese, English and Russian are identified respectively from others. Therefore, automatic identification of four scripts is realized successfully.