onaversatileandcostlessomrsystem.docVIP

On a versatile and costless OMR system IOSIF I. ANDROULIDAKIS, MSc Network Operations Center University of Ioannina GREECE NIKOLAOS I. ANDROULIDAKIS, MSc in Management Nottingham Trent University  Abstract: In the present work is presented a versatile and costless Optical Mark Recognition system. It is used to automate the import into a pc of data that emanate from multiple choice questionnaires. This becomes feasible by using a scanner providing speed and reliability advantages and at the same time ease of use against the traditional by hand method of import. Its main advantage is the ability to work with every single scanner of the market, needing no extra drivers whatsoever and no special colour printing of the forms. Furthermore, the user doesn’t have to mark manually the areas of interest as is done in most other OMR implementations. The program is also small and compact, easily fitting in a diskette. Following the introduction and the analysis, we are trying to give a clear analysis of the optical recognition algorithms that the program uses. Later the way of use is presented, accompanied by figures. Results from its pilot use follow as also an evaluation benchmarking comparing the traditional way of data import. It is shown that the time is decreased at least an order of size. Finally, ideas for the application’s future expansion are mentioned which already has begun in order to treat numbers and more complicated forms of data. Key-Words: OMR, OCR, pattern recognition, scanner, questionnaire, gallop 1 Introduction The size of the pool in a poll is one of the most important factors in order to get objective results [1]. However the import of questionnaire’s data in the computer for further analysis, is a tedious, precarious and boring process, especially when the number of questionnaires increases. Our solution can automate this process providing speed, reliability and ease of use against the by hand import method. It allows the automatic imp