Effect of supervised learning methodologies in offline handwritten Thai character recognition

Abstract

Offline handwritten character recognition is a conversion process of handwriting into machine-encoded text and predominantly used for digitizing handwritten texts and forensic applications. Currently, several techniques and methods are proposed to enhance accuracy of offline handwritten character recognition for many languages spoken across the globe like English, Tamil, Chinese and Arabic. In this paper, a local feature-based approach using supervised learning techniques is proposed to enhance the accuracy of handwritten offline character recognition for Thai alphabets using unsupervised learning for individual character as a class, whereas most of the existing methodologies for Thai character recognition is done with group of similarly looking characters as a class. The classification is operated by using support vector machine (SVM). The accuracy would be the percentage of correct classification for each class. For the result, the highest accuracy is 74.32% which has 144-bit shape features and uniform pattern LBP for the features.