Continuing our investigations into authorship verification in paintings, we are happy to present our newly accepted results in the IEEE Symposium Series on Computational Intelligence (IEEE SSCI 2021). We now work with 79,432 paintings, 1584 painters, 135 styles, and 42 genres 😍❤️
Connoisseur: Provenance Analysis in Paintings by Lucas David
Helio Pedrini , Zanoni Dias and Anderson Rocha (Artificial Intelligence Lab., Recod.ai , http://Recod.ai ) at the Instituto de Computação – Unicamp .
In this work, we employ convolutional network-based strategies to identify and classify art-related digital artifacts over the Painter by Numbers dataset. We propose to exploit the authorship, style and genre annotated information in a multi-task setup, in which patches of paintings are encoded through a multiple outputs network and, in a second stage, used in a Siamese discriminating network to solve the authorship matching problem. We also combine the available annotated information in a more efficient manner, by posing the Painter by Numbers challenge as a multi-label problem.