Abstract:
One of the main factors in data collection and processing is using computers and
scanners as manageable and instantaneous tools, to make massive archives. Nowadays,
questionnaire forms with multiple choice boxes can only be marked with pencil, and
then by using special scanners, the system detects the carbon marked boxes for data
processing. The designed system is a software package that prepares a report which is
based on scanning a regular questionnaire form and detecting the marked boxes. Its
detection is independent of pencil or any other dark writing pen. In addition, the boxes
do not need to be filled in entirely, meaning that they can even be marked with an "X"
or even a check mark. Every sheet is scanned by a typical fast scanner with an
Automatic Feeder Device (AFD) and a minimum 200 (Dot Per Inch) DPI. After
scanning the sheets in gray scale level and JPG format, each one will be saved as a
picture file into a folder. Then the folder will be extracted and the files will be converted
into binary as black and white according to a predefined algorithm. This algorithm
opens the folder again and starts searching the marked area's region, if the filling area
percentage is over 40%, it will be counted as marked. Another feature of this system is
its capability to detect and correct the rotation of the sheet up to 30 degrees. After the
correction is made, the sheets will be scanned and the marked locations are saved into
the database for further report extraction. This system was tested with more than 10,000
survey sheets filled in by students. The scanned sheets were compared with the originals
and the error deviation was less than 0.04 percent which the failure was due to badly
(less than 40% filled) marked boxes and jammed paper. Semnan University is currently
using this system for evaluating faculty teaching performance.