Printed Uyghur Word Recognition

Character recognition is an important field of the pattern recognition applications. Nowadays, compared to English, French and other mainstream languages, the research on Uyghur recognition technology is relatively less and not very mature. However, research about Uyghur recognition technology own great requirement.In addition, it is very important to develop our national culture. The paper focuses on printed Uyghur word recognition based on research on Uyghur word and printed character recognition. The main contents are as follows.1. Introduced research background, significance and research status on the printed word recognition. In addition,printed Uyghur word characteristics have been introduced.2. Detailed analysis about the preprocessing of printed Uyghur word recognition. Pretreatment section includes: binarization, noise elimination process, layout analysis, tilt correction, normalization and contour processing. Therefore, it can keep the most features of printed Uyghur word to extract features and classfy.3. Further studies about printed Uyghur word feature extraction and classification process. Feature Extraction: the 4-direction line element, gradient features and Gabor features. Classifier using Euclidean distance classifier and BP neural network classifier classifier.4. Realized Android-based Uyghur single character handwriting recognition software. The JNI technology is used to utilizing this software.