An object-oriented expert system for Thai character recognition

Character recognition is one of the most useful computer applications. The progress in this area contributes to the progress of information systems since data entry is an accepted bottleneck in information system operations. Thai character recognition is also an active research topic. Various researchers work on the recognition of both printed and hand-written Thai characters which are unique and different from English alphabets. A popular classification technique for Thai characters is the use of structural characteristics such as concentrated code values (Q-codes). Nowadays, the creations of rules which are used to classify characters are done manually which may result in long, redundant and incomplete rules. This research work presents the development of an expert system for Thai character recognition. It accepts training sets and produces a concise and complete decision tree which gives classification rules for Thai characters. The system was developed using the popular object-oriented programming language Smalltalk