Handwritten Manuscript Digitizer

In India, there are various instances where information is gathered by filling a questionnaire or a form. This information is then updated manually into the databases by the concerned authorities. Due to manual data entry, human error results in the capture of inaccurate data and thereby results in faulty storage and analysis of the data. The process is time consuming with a greater probability of error. This document serves as a guideline to automate and expedite the above process. The paper contains ideas of converting the handwritten samples into electronic data. It uses the kernel method of Multi class Support Vector Machine for handwritten character recognition. The data is first extracted in form of individual images for the corresponding data field, pre processed and converted to digital format. This reduces the time and human effort needed for the same. This paper aims at easing the process of evaluation by automating the correction process.