Thai OCR: a neural network application

Thai optical character recognition (Thai OCR) is one of the most desirable computer applications in Thailand at present. Though there are several proposed techniques for solving the problem, none seems to produce a satisfactory practical result. Many limited factors are encountered such as the inconsistency of the scanning process, the incomplete and/or noisy original documents, and the shift and position variance of the recognition technique. We proposed to apply artificial neural networks (ANNs) together with some pre-processing and post-processing techniques to solve the Thai OCR problem. The experimental results confirm that ANNs are a very suitable technique for developing the Thai OCR software. The recognition rate on a real document of training fonts is about 90%-95%. This leads to a possible implementation in a production-quality OCR software that the NECTEC software technology laboratory is working on. The details of all processes are explained in the paper.