An improved handwritten Chinese character recognition based on Localized Ellipse model

A feature selection technique along with a Localized Model procedure for improving the recognition accuracy of a visual-image-based handwritten Chinese character recognition system is presented in this paper. A novel Localized Ellipse Model approach is developed and implemented on the feature extraction from the visual Chinese character. Smaller sub-regions from a predefined neighborhood within the Ellipse Model of the training samples are merged to obtain a large set of features. These features are then projected into the Eigen Spaces using PCA (Principal Component Analysis) methods. At the end, using the proposed SVM based binary tree, character is classified as an original or a forgery one. The recognition system has achieved a high recognition rate of 85% on SWU database, a handwritten Chinese character database.