A Microservices-Based Optical Mark Recognition Framework with Advanced Image Processing and Cloud-Native Architecture for Scalable Educational Assessment

This article presents a comprehensive LaTeX manuscript detailing the design, implementation, and evaluation of an automated Optical Mark Recognition (OMR) system for efficient and accurate multiple-choice question (MCQ) assessment. The system integrates advanced computer vision techniques, a user-friendly web interface, cloud-ready deployment strategies, and robust evaluation pipelines to minimize manual grading efforts, enhance consistency, and enable scalable academic workflows. By leveraging image preprocessing, template matching, and contour analysis, the OMR system accurately detects marked bubbles on scanned answer sheets, supporting both single and multiple-choice formats while incorporating quality assurance mechanisms. The methodology outlines a modular architecture encompassing image acquisition, preprocessing, answer recognition, evaluation, and reporting, with emphasis on security, auditability, and modularity. Experimental validation includes architecture and deployment diagrams, demonstrating high recognition accuracy (98\%) and efficient processing (1.3 seconds per sheet). Results are presented with performance tables and scalability diagrams, showing reliable operation under varying loads. The literature review integrates 41 references from the author's bibliography, covering intelligent systems, cloud applications, AI optimization, and educational technologies. Conclusion and discussion highlight the system's benefits for academic institutions, while future work explores AI enhancements, federated learning, and blockchain integration. This work contributes to the field of automated educational assessment by providing a complete, referenced manuscript that bridges traditional paper-based exams with modern digital evaluation paradigms.