Development of an effective system for real-life applications essentially require its training and testing on a reasonably large database of real-life samples. However, to the best of our knowledge, NIST Special Database 2 (SFRS) and NIST Special Database 6 (SFRS2) are the only two available standard form databases and these consist of simulated form data.


The present database include scanned images of 1950 filled-in application forms for the admission test of the year 2004-05 of Indian Statistical Institute, a premier academic Institute of India. It also includes 25 filled-in application forms for each of (i) Indian Passport, (ii) BSNL mobile connection and (iii) Union Public Service Commission examination of India. The last three categories of application forms are filled-in by students who worked in our laboratory at different occasions. The samples of the first category include form images with poor quality impression due to photocopying or printing using low quality printer and/or ink. The samples of this database were scanned using an HP flatbed scanner at 300 dpi and stored in the grayscale tif format.


Back to Ujjwal's main page