Abstract:
Multiple-choice questions are a widely-preferred tool used to evaluate learner’s academic
performances in education institutions all over the world. However, in situations when the
number of individuals is large, this exercise becomes a monotonous, tedious, and time consuming piece of work that is susceptible to errors. Due to the mistakes committed by
instructors while marking, students are awarded wrong grades.
This project was therefore aimed at developing a mobile application that would solve the above
problem by enabling instructors scan the reference/ key (marking guide) answer sheet once
which is stored in the database and then take pictures of students’ handwritten-on answer
sheets which are used to extract characters, process and recognize them by the OCR and OCR
engines. These extracted characters are cross referenced by the marking algorithm to mark the
students’ performance hence saving lecturers’ valuable instruction time.
The work is arranged mainly in six chapters, chapter one includes the introduction of automatic
multiple-choice answer sheet marking mobile application. Chapter two discusses the literature
related to the system, chapter three illustrates the methodologies used in coming up with the
working prototype of the system, chapter four includes system design and analysis, chapter five
contains the implementation and testing of the system and chapter six contains the summary of
the work, challenges, critical analysis, proposals and recommendations.
