Unimore logo AImageLab
Back to the project list

YNAP Artificial Intelligence Lab@UNIMORE

YNAP Artificial Intelligence Lab has been established in conjuction and with the support of Yoox Net-A-Porter Group. Within the YNAP AI Lab, AImageLab will study and develop different technologies of Computer Vision and Artificial Intelligenge applied to the fashion domain. 

As a leading company in the use of new technologies, Yoox Net-A-Porter Group, the most advanced fashion hub in Italy for Artificial Intelligence and Visual Recognition, has decided to join forces with the University of Modena and Reggio Emilia and has created a joint research program dedicated to Artificial Intelligence and Computer Vision in the field of fashion.

The three-year partnership with AlmageLab has given birth to the research program led by Professor Rita Cucchiara. The final aim is to completely transform the online shopping experiences of customers around the world, developing cutting-edge technological solutions thanks to the processing of the enormous archive of the Yoox Net-A-Porter Group, which contains over 20 years of data and tens of millions of images.

As a starting point, the collaboration is mainly focused on two research topics: Visual Search and Virtual Try-On. The solutions studied will allow to identify in the images of the Yoox Net-A-Porter database not only the garments but also some specific features such as materials, lengths, colors, and classify them according to these traits. Moreover, additional functions based on artificial intelligence will be developed to further enhance the user virtual experience of interacting with clothing items.

"Collaboration with one of the world's leading experts in the study of Artificial Intelligence makes it possible to explore the potential of technology to support humanity innovatively".
- Federico Marchetti, President and CEO of YOOX NET-A-PORTER GROUP

"YNAP AI LAB@UNIMORE is an excellent context to test scientific results from the analysis and study of images, on which we have always worked. This collaboration will allow us to maximize the efficiency of YOOX NET-A-PORTER's internal systems, thanks to the retrieval of useful information from images, using the latest generation techniques, and to go even further, using innovative approaches, such as the Virtual Try-On, to promote increasingly unique shopping experiences for customers".
- Rita Cucchiara, Scientific Director


Visual Search

The goal is to find the most similar garments in the YNAP database starting from a user-generated picture. This problem involves different computer vision tasks: the identification of all garments in the picture, the classification of the garments, and the retrieval of the most similar items from a pre-defined database. In this context, AImagelab will be mainly focused on the development of different deep learning-based solutions to address the retrieval task.


Virtual Try-On

This task will involve the development of new generative neural networks capable of creating new views of a specific garment or swapping the garments in an outfit with other garments.



1 Fincato, Matteo; Cornia, Marcella; Landi, Federico; Cesari, Fabio; Cucchiara, Rita "Transform, Warp, and Dress: A New Transformation-Guided Model for Virtual Try-On" ACM TRANSACTIONS ON MULTIMEDIA COMPUTING, COMMUNICATIONS AND APPLICATIONS, vol. 18, pp. 1 -23 , 2022 | DOI: 10.1145/3491226 Journal
2 Fincato, Matteo; Landi, Federico; Cornia, Marcella; Cesari, Fabio; Cucchiara, Rita "VITON-GT: An Image-based Virtual Try-On Model with Geometric Transformations" Proceedings of the 25th International Conference on Pattern Recognition, Milan, Italy, pp. 7669 -7676 , 10-15 January 2021, 2021 | DOI: 10.1109/ICPR48806.2021.9412052 Conference
3 Morelli, Davide; Cornia, Marcella; Cucchiara, Rita "FashionSearch++: Improving Consumer-to-Shop Clothes Retrieval with Hard Negatives" Proceedings of the 11th Italian Information Retrieval Workshop, IIR 2021, vol. 2947, Bari, Italy, September 13-15, 2021, 2021 Conference

Video Demo

Project Info



01/11/2019 - 31/10/2022

Funded by:

Yoox Net-A-Porter

Project type:

Joint Research Program