Happy Finish, the VR content specialists and global production studio, today announces the launch of its Artificial Intelligence (AI) technology. Shoegazer, currently a proof of concept (PoC) for the fashion-retail industry, uses image recognition and transfer learning technologies to identify brands and models of trainers in real-time – with 95% accuracy.
Happy Finish’s Shoegazer is a unique PoC for the fashion and retail industry
The current prototype uses image recognition to identify certain features of trainers – including logo to shape and style. By using transfer learning technology, the app is able to match the picture a consumer takes of a trainer on their smartphone with pre-existing data sets, identifying the exact brand and style of the trainer in real-time. Using two comparative tests for benchmarking, one based on accuracy and the other on computational time, Shoegazer identifies trainers with 95% accuracy.
There is also opportunity to apply the transfer learning technology from Shoegazer to other areas within fashion and retail, including clothes, accessories and bags. The nature of the application also means that it is not solely limited to individual brands – there is also an opportunity for retailers to use the application to cross-sell other products within their portfolio. The prototype is already being presented to partner brands with a view to market launch in late 2016.