The main purpose of the Project is to develop an innovative tool – an automated engine which generates recommendations in the form of fashion stylisations tailored to the user. The stylisations are to be based on advanced artificial intelligence and machine learning algorithms that analyse and model fashion styles and patterns on massive data sets. The final solution will be available both as:
1. a mobile application Personal fashion stylist (targeted at individual users; selection of clients' own clothes and clothes from stores, i.e. institutional clients - the Applicant's clients),
2. an advanced tool-module for online clothing stores (selection based on the store's assortment, but the model itself trained on global data).
The Project creating process
Expected results
As part of the Project, new algorithms and methods of artificial intelligence will be developed in the field of analysis and modeling of fashion styles. Computer vision and machine learning methods will be used, in particular methods based on deep neural networks, as well as methods of studying the relationships of complex objects with the use of similarity in the form of, for example, a network of comparators. Instead of creating precise style definitions and manually defining visual attributes that define the style of clothing, the Company will rely on discovering knowledge about fashion styles from large data sets.
The result of the Project will enable the e-commerce industry and individual users to create personalised recommendations of fashion stylisations – sets matched to each other in relation to the elements of clothing and accessories composed of the available assortment, e.g. a store's assortment or a user's wardrobe.