eTryOn developed 3 novel XR apps: i) the VR designer app, a desktop VR application targeting fashion designers, offering immersive and realistic visualizations of their digitally designed garments and how they fit on animated 3D avatars, ii) the Dress Me Up app, a web application for social media users (e.g. influencers), allowing them to virtually change their outfit on a photo by selecting from a pool of digital garments and then upload it to social media, and iii) the Magic Mirror app, a mobile-based AR magic mirror enabling virtual tryons of garments during online shopping that aims to recreate at home the experience of buying clothes from a physical store.
In creating these 3 apps and their components, we have developed a body self-scanning software with significant improvements in the accuracy of the creation of 3D body models, that can produce a dressable 3D avatar for use in the Dress Me Up application. Afterwards, relying on real garments from past, ongoing and new collections, we created digital assets that can be used in game engines, and are subsequently used in the VR Designer and Magic Mirror apps that run on the Unity game engine. Regarding the simulation of the developed avatars and the selected garments, the transformations of the aforementioned garments in forms compatible for simulation in game engines and AR frameworks were defined. For extracting fashion insights, we have developed a Machine Learning model for the detection of categories and attributes within fashion imagery. Furthermore, the aforementioned model in combination with past trend data were used to form a fashion trend analysis and prediction model. On combining all the aforementioned technologies in standalone XR apps, we have set up a cloud-native, serverless-first architecture design delivered and iteratively kept up-to-date as the project evolved and incorporated an Architectural Decision Records methodology. Furthermore, detailed technical diagrams have been delivered with clear distinctions of the cloud functions that were implemented that aided all three XR apps. Regarding the development of the applications, we firstly analysed the options and selected the platforms for the software implementation of each app. All the aforementioned technical work relied on the detailed user requirements that were gathered through questionnaires directed to the target audience of each app. Cloud function integration, API implementations and the development of the three apps has been finalized. Furthermore we defined a clear plan of testing the applications in our pilot sessions and carried out several dissemination activities on various social media channels to promote said sessions. We carried out the three different pilots and we got feedback from our end users, in the form of questionnaires (subjective) as well as from analytics services that we have set up (objective).
