Validation of User-submitted Images Using Machine Vision

Abstract

The current manual validation of digital passport-sized photographs submitted to various institutions is subjective and inefficient. This paper addresses the need for an automated system to validate the quality of these photographs, as the existing systems only check image format and file size. The proposed solution leverages machine vision techniques to assess composition, clarity, background, head position, brightness, sharpness, and blurriness against predefined criteria. This AI-driven approach promises to enhance the scalability, accuracy, and efficiency of photograph validation processes, particularly for organizations like the Kenya National Examinations and telecommunication companies in Kenya, which handle millions of images.