Image Processing-Based Handwriting Recognition for Automated Form Processing

Processing forms need a substantial amount of work and time, especially when the human resource is lacking to store it manually into the system. A method is proposed in this study to develop the handwriting recognition for automated form processing. Handwriting of uppercase and lowercase letters, together with numbers from 0 to 9 are collected using Form A, a form designed for it. The handwriting on the form will be pre-processed using image processing techniques and binary sub-images of the handwriting should be obtained. Letters and numbers are kept as templates in MAT file. Form B is a form designed based on the Universiti Malaysia Sabah (UMS) Course Registration Form to be filled in and compared with the template. Form B will undergo the same processes as Form A to get the same format as the final output of Form B. For the evaluation of recognition rate, ‘normxcorr2’ function of Matlab is used. The applicant’s information is extracted to undergo a series of more steps. These individual numbers and letters will be compared with the template to obtain the result in text form. Graphical User Interface (GUI) in Matlab is used to present the system. The result of the post processing is the handwriting which is displayed in text form in a Notepad file and the applicant’s information is stored in a Microsoft Excel file.