Aistetic is an Oxford University spinout providing accurate digital remote measurement, sizing and avatar creation using the latest AI-based 3D reconstruction techniques from data captured via smartphone cameras. Aistetic's current product enables retailers in the $821B/£652B, global fashion e-commerce market to match the sizing of ready-to-wear or tailored fashion items to the shape and size profiles of customers, improving customer satisfaction and reducing return rates for clothing purchased online. **PrelovedAI** is a collaboration between Aistetic and fashion reseller Stuff-u-Sell to undertake feasibility research and first-stage technology development for a new AI-based capability to digitally verify the quality and condition of preloved or returned fashion items for resale. The technology will unlock _business productivity_ through the automated processing of manual administrative tasks using AI to deliver highly-accurate quality classifications from photo imagery. The technology will reduce costs and increase capacity for online fashion retailers across two large and growing markets - Preloved/Secondhand fashion resale, ($197B/£157B global value), and fashion returns, (with 27% of all items purchased online returned). The project will innovate by leveraging Aistetic's expertise in deep learning-based AI methods to develop an AI framework that is trustworthy and can be easily scaled across clothing types, fabric, patterns, defects, and input image quality. The project will deliver a working AI-model and evaluation dataset. From which solution feasibility, accuracy, and productivity improvements can be measured. It will improve trust for resellers and increase the likelihood that clothing can be resold, rather than being wasted or destroyed.

no public description

Aistetic is a unique business proposition bringing together the leading experts in Deep Learning based Artificial Intelligence & Computer Vision research, combined with a huge £265bn market opportunity to disrupt the traditional e-commerce model for buying and selling of Men's tailored fashion. Duncan McKay has significant experience developing and bringing innovation to market having launched a new juice brand delivering £18m in retail sales in the first 12 months of launch, led the Quaker business in the UK (£150m retail sales business) and has been Head of eCommerce for the Pepsi Lipton European Business Unit. Professor Phil Torr has a world leading track record in computer vision research and taking technology to market (Oxsight, Five AI, 2D3, and Sony's Wonderbook development for PS3). We are in the early stages of developing our innovation and with this grant funding we intend to leverage deep learning driven AI and image based 3D reconstruction, and draw upon the latest scientific advances in statistical human shape analysis to take accurate measurements of your body easily and quickly using just your mobile phone under diverse conditions. We will first sell made to measure shirts online through our new app before expanding our offering into other tailored Men's garmets, including jackets and trousers. This novel software technology for the mobile phone addresses a number of problems with current solutions: (1) Measurement and inaccuracy. (2) Too many inputs (data, physical accessories) -- individuals are required to strip down to very tight clothing, provide a lot of body data, or wear a yoga suit. Grant funding will provide the means to develop a first working prototype of the Aistetic app featuring the core deep learning AI innovations. The collaboration will enable meaningful and tangible commercial output to research work undertaken by the University of Oxford.