Script to evaluate OMR's this Shaastra Ignite.
Outputs your marked answers.
Opencv3
Python 2.5+, 3.4+
python omr.py input_image
All intermediaries will be visible as separate panes.
Use any key to exit.
-
Firstly we obtain a perspective transform, since we have a layout of the OMR sheet at hand
-
Proceeding this we isolate areas where OMR bubbles are marked. This is done via edge detection and filter usage.
-
Lastly, we create maps to isolate bubbles row wise, and compare their thresholded pixel values.
- Add CRNN integration to detect name, roll no..etc