iSUKAT: Capture, Measure, and Know Your Perfect Shoe Size Using Image Recognition

iSUKAT represents a groundbreaking web-based application that harnesses advanced image recognition technology, specifically Roboflow 3.0 Instance Segmentation with a COCO-seg checkpoint, to revolutionize foot measurement and shoe selection. By employing sophisticated machine learning algorithms, the system accurately captures and analyzes foot dimensions, identifying foot types and shoe brands with high precision. The application directly addresses the critical issue of improper shoe sizing, which impacts 63-72% of consumers and leads to significant discomfort and potential foot health problems. Utilizing a comprehensive evaluation methodology based on ISO 25010 quality standards, the research assessed the system's performance through quantitative metrics including Mean Average Precision for foot detection and segmentation, instance segmentation accuracy, and real-time image processing inference time. The empirical results were notably impressive, with the system achieving excellent ratings across key performance domains: Functional Appropriateness scored 3.35/4.00, Usability reached 3.26/4.00, and Reliability was evaluated at 3.31/4.00, resulting in an overall system acceptance of 3.21/4.00, which was classified as satisfactory. The application strategically addresses e-commerce shoe shopping challenges by providing personalized recommendations based on precise foot measurements. However, the research also revealed nuanced performance insights, with only 11.6% of users strongly agreeing on measurement precision, while 40.5% expressed concerns about measurement accuracy. Despite these challenges, the study ultimately underscores iSUKAT's transformative potential in digital footwear selection, presenting an innovative solution that seamlessly integrates advanced image recognition, machine learning, and user-centric design to dramatically enhance the online shoe shopping experience.