An intelligent character recognition system for automatic mark capturing

Rising admissions in the South African institutions of higher education have enlarged student-to-lecturer ratios and increased the lecturer´s workload, already burdened by administrative tasks. After marking tests, lecturers usually fill in a document called the cover page where the student´s number, name and marks according to the questions are placed. Once this is done, they will have to recopy those marks on a spreadsheet for record purposes. This research aims to recognize the handwritten characters (student numbers and marks) on the coversheet as well as to automatically allocate the marks on the spreadsheet according to students´ numbers using Matlab. The Feedforward Backpropagation Neural Networks and the Matlab are used to recognize handwritten characters and communicate with Microsoft Excel.