Recognition of Offline Handwritten Chinese Characters Using the Tesseract Open Source OCR Engine
Qi Li, Weihua An, Anmi Zhou, Lehui Ma · 2016
Due to the complex structure and handwritten deformation, the offline handwritten Chinese characters recognition has been one of the most challenging problems. In this paper, an offline handwritten Chinese character recognition tool has been developed based on the Tesseract open source OCR engine. The tool mainly contributes on the following two points: First, a handwritten Chinese character features library is generated, which is independent of a specific user's writing style, Second, by preprocessing the input image and adjusting the Tesseract engine, multiple candidate recognition results are output based on weight ranking. The recognition accuracy rate of this tool is above 88% for both known user test set and unknown user test set. It has shown that the Tesseract engine is feasible for offline handwritten Chinese character recognition to a certain degree.