Answer Sheet Recognition System Based on MATLAB [GUI Interface + Thesis]

The research content of this project is as follows: The overall scheme of the answer sheet recognition system was established. This scheme considers not only the system framework and hardware composition but also includes software development plans, with primary emphasis on software implementation to ensure efficient system operation. Digital image processing algorithms were determined. Image preprocessing techniques including tilt/rotation correction and binary transformation were implemented. After analyzing and comparing various processing methods considering image characteristics, Hough transform was selected for detecting line inclination angles in images, followed by image rotation for correction using MATLAB's image processing toolbox functions. Methods for identifying answer-marking areas and student ID regions using MATLAB software were developed. Successful implementation included accurate extraction of student IDs and answers through region segmentation algorithms, effectively handling both multiple-choice and single-choice questions. The software component programming was completed using MATLAB. Primary functions involved extracting marked answers from answer sheets and comparing them with standard answers. Through computational algorithms, the system calculates correctly answered questions and computes final scores. Statistical scores were compared with actual results, with analysis conducted to identify discrepancies. This step helps evaluate system performance and identify potential improvement areas through error analysis routines. The above constitutes the research content of this project. Continued efforts will focus on system refinement to provide enhanced service through algorithm optimization and validation testing.