A Novel Japanese Character Recognition Algorithm for Automatic Smart Courseware Generation Tasks

The efficient recognition and smart analysis of the character is essential for the digital information processing. To deal with the traditional algorithms of the low efficiency or noise-sensitive features, this paper propoes novel Japanese character recognition algorithm for the automatic smart courseware generation tasks. The proposed algorithm has 3 aspects of considerations: 1) The traditional ELM is optimized considering the unsupervised features, to then obtain the fast deep learning network training speed; 2) The proposed OCR method is studied, the pre-processing steps are combined to make the framework complete; 3) The automatic smart courseware generation scheme is integrated as the application. In the experimentation section, the proposed model is compared with the SVM and PCA, the results on accuracy has proven that proposed model is efficient.