Cyber-Trust

Advanced Cyber-Threat Intelligence, Detection, and Mitigation Platform for a Trusted Internet of Things

Financed by

European Commission – H2020

Duration

01/01/2015 – 31/12/2017

Total project cost

EUR 4.084.845

DaVinciFace Project

Financed by: Mathema SRL
Type: Project
DaVinciFace is a software - based on the most innovative techniques of Artificial Intelligence, in particular on GAN (Generative Adversarial Network) - which aims to capture Leonardo's pictorial style and to attribute it to faces, mainly female, to create portraits in the Leonardesque style ( eg La Belle Ferroniére, the Mona Lisa).

How it works

The software is made available to users through an app, through which the user can take a photo of the face (selfie) and upload it to the Mathema server where the photo is pre-processed, so that the dimensions are compatible. with those of the final elaboration, and subsequently elaborated with Deep Learning techniques performed on models trained on suitable datasets of faces and portraits. A video is then returned to the user, showing the transition from the original face to the digital portrait, with the acquired Leonardo characteristics, so that the user can select the frame that best suits him (from a greater similarity to the original subject to the greater correspondence to a pictorial work) and then obtain your digital portrait in Leonardesque style in high resolution (300 dpi). This portrait can later be used for a physical copy, eg. print on canvas to get a picture.