DaVinciFace generates portraits in the style of Leonardo da Vinci from an uploaded photograph, using a deep learning model built specifically for that single transformation.
Committing to one artist is what makes it work. General style transfer applied to a famous painter usually produces something that gestures at the reference – the palette roughly right, the handling wrong – because the model is spreading its capacity across every style at once. A model built for one body of work can capture the specific things that make it recognisable: the sfumato, the particular treatment of light on a face, the three-quarter turn.
Prints are sold alongside the digital output, which suits what people actually want from a portrait like this. It is deliberately narrow – one style, one purpose, no editing controls – so judge it as a keepsake rather than a creative tool, and take the vendor’s “the only AI that” phrasing as marketing rather than a verified claim.
