Personalised Size Recommendation for Online Fashion

The aim of this project is to develop and test the garment size recommendation app, ShapeMate, embedded within a fashion e-commerce site. Not finding the correct garment size causes the high return rate of 30%-40%, in fashion e-commerce. The app captures a single image with minimal user input, to estimate and classify the 3D body shape, in order to generate body measurements and using this information, to match with garment data for size recommendation. An extensive user-experience study was conducted, which was empirically tested through semistructured focus group interview and survey, to validate results and obtain further insight. This research project offers a major innovation for low-cost size recommendation generated from a single image for fashion e-commerce, enhancing the online apparel shopping experience, by matching personalised body measurements with customised garments.