Automated OMR Analyser using ML and Image Processing

This work explores Optical Mark Recognition (OMR) systems built on Machine Learning and Image Processing techniques to achieve accurate and automated evaluation of answer sheets. The primary objective is to eliminate dependency on specialized scanners by utilizing devices such as mobile cameras and webcams. This system processes scanned answer sheets by isolating responses and matching them with reference answers using automated logic. Several models integrate contour detection, circle detection, and error-tolerant classification to ensure robustness even with imperfect markings. The survey shows that OMR systems deliver high accuracy in diverse lighting and marking conditions. The findings support the development of scalable and cost-effective OMR solutions that are particularly useful in academic and institutional settings.