OMR Sheet Analysing Using Computer Vision
Sai Charan Goli, Sai Teja Gattu, Vishnu Vardhan Goud Kalal, P. Michael Preetam Raj, S. Karthick, Sravan Kumar Vittapu · 2025
For years, Optical Mark Recognition (OMR) have utilized to extract data through paper-based forms. However, analyzing OMR sheets by hand is time-consuming and prone to human error. This study introduces a sophisticated method which combines advance image processing and neural learning techniques to systematize OMR sheet breakdown. The proposed design exactly identifies and marks responses on the bubbled sheets through the application of image enhancement, segmentation, and feature extraction algorithms. To assess individual and group performance, a performance analysis is conducted, offering valuable insights for educators and researchers. Various imaging methods, including Median filtering, RGB to Grey conversion, edge detection, and complement of image, are employed to evaluate the sheet and display the final score based on the selected responses for each question. This approach eliminates the need for bulky equipment and expensive scanners, resulting in time and cost savings. Advanced algorithms and preprocessing techniques are used to address complex issues related to noisy image, marking variations, alignment of sheets Through extensive experimentation besides comparison, the system validates its greater accuracy and effectiveness likened to method of manual analysis. The paper also explores potential future improvements, such as cloud-based combination and instantaneous analysis, prominence the system's capacity aimed at greater automation and reliability in valuation process.