From Stroke to Finite Automata: An Offline Recognition Approach

Kehinde Aruleba

Abstract: A major challenge in making online education easier and more effective lies in developing automatic recognition, interpretation, and grading systems that can provide meaningful feedback to lecturers and students. Formal Languages and Automata Theory is a major module for many computer science and computing programmes worldwide. In such a module, students are taught how to design a finite state machine to recognise words in any given language. Despite the wide acceptance of this module by most universities across the globe, most students find this module difficult, boring and too abstract. Several research has been conducted on how to make this module interesting, but there still exist some gap. In this work, we propose and implement a system that assistant learners in learning this module. The system is in two units: unit one focus on offline recognition of hand-drawn finite automata diagram and the second unit focus on the tutor system. The system produced a 97% recognition rate. In future work, we intend to use formal grammars (second unit) to represent the recognised FA components, this will be used to automatically parse the output of the recognition system to determine if valid FA has been drawn.