DeepAI
Log In Sign Up

Raiders of the Lost Art

09/10/2019
by   Anthony Bourached, et al.
UCL
0

Neural style transfer, first proposed by Gatys et al. (2015), can be used to create novel artistic work through rendering a content image in the form of a style image. We present a novel method of reconstructing lost artwork, by applying neural style transfer to x-radiographs of artwork with secondary interior artwork beneath a primary exterior, so as to reconstruct lost artwork. Finally we reflect on AI art exhibitions and discuss the social, cultural, ethical, and philosophical impact of these technical innovations.

READ FULL TEXT VIEW PDF

page 1

page 2

page 3

page 4

06/03/2019

Depth-Preserving Real-Time Arbitrary Style Transfer

Style transfer is the process of rendering one image with some content i...
05/28/2018

A Pragmatic AI Approach to Creating Artistic Visual Variations by Neural Style Transfer

On a constant quest for inspiration, designers can become more effective...
08/09/2022

HyperNST: Hyper-Networks for Neural Style Transfer

We present HyperNST; a neural style transfer (NST) technique for the art...
05/26/2016

DeepMovie: Using Optical Flow and Deep Neural Networks to Stylize Movies

A recent paper by Gatys et al. describes a method for rendering an image...
02/13/2021

Hiding Data Hiding

Data hiding is referred to as the art of hiding secret data into a digit...
07/31/2020

Neural Style Transfer for Remote Sensing

The well-known technique outlined in the paper of Leon A. Gatys et al., ...

1 Introduction

The beauty of a piece of art is often the insight it gives the viewer into the vision of the creative mind behind it. Visual pieces are an aesthetic bridge into the artist’s inspiration and perspective. This bridge does not make a one to one mapping and so conclusions are always non-unique and the meaning always remains subjective. While the primary exterior is the result with which the artist was pleased, it does not show their failed attempts; the diffuse musings of their creative process.


Figure 1: Visual method of reconstructing the art hidden beneath The Old Guitarist pablo_guitarist pablo_la_vie .

Here we attempt to broaden the insight into an artist’s intentions, mistakes, and musings by reconstructing artwork that has been hidden under primary exterior of some of the most inspirational artists in history. We believe that this gives one of many possible inferences of what inspiration existed in the artist’s mind.

2 Methods

In this study we considered two paintings; The Old Guitarist (1903-1904), and The Crouching Beggar (1902), both painted by Pablo Picasso. Recently, art conservationists have used techniques in spectroscopy and radiography to reveal hidden art beneath a primary exterior piece. Concealed beneath The Old Guitarist hides the portrait of a woman and below The Crouching Beggar a mountainous landscape. Through x-radiography these underlying pieces have been made apparent xray .

In this study we combined the techniques of neural style transfer gatys and x-radiography. We edited the x-radiographed artworks to remove the impression of the primary exterior artworks. This process involved manually removing features that were deemed unlikely to be present in the interior secondary pieces. For example, The Old Guitarist’s eye is highly unlikely to feature on the left shoulder of the woman in the secondary interior piece.


Figure 2: Visual method of reconstructing the artwork hidden beneath The Crouching Beggar pablo_beggar pablo_terraced .

Using the manually edited x-radiographed image of The Old Guitarist as the content image and La Vie as the style image, we applied neural style transfer to reconstruct the lost artwork; creating the lost woman, La femme perdue. La Vie was chosen as this painting was produced pre-1904, during Picasso’s Blue Period, and includes mutually aesthetic features. A similar process was applied to Picasso’s The Crouching Beggar. Beneath The Crouching Beggar lies a mountainous landscape, that art curators believe is of Parc del Laberint d’Horta, near Barcelona. They know that Santiago Rusiñol painted Parc del Laberint d’Horta. Santiago Rusiñol’s Terraced Garden in Mallorca was chosen as the style image as its style and content appeared similar to that of the secondary interior piece of The Crouching Beggar. Using a manually edited x-ray fluorescence image of The Crouching Beggar as the content image and Santiago Rusiñol’s Terraced Garden in Mallorca as the style image, we applied neural style transfer to reconstruct the lost artwork; creating Rusiñol’s Parc del Laberint d’Horta.

3 Discussion

From multiple art exhibitions, where we were the only presenters of AI art, we discovered that there is often skepticism that an algorithm can be creative, or innovate. We propose that this skepticism is rooted in the belief that a human-like, non-precise, abstract concept can not be represented in a machine. Humans’ ability to empathise with something that is very similar to it, like other humans and animals, but not machines is likely the origin of this skepticism 111We call this bias the Empathy Paradox.

Deep convolutional neural networks enable the representation of abstract visual concepts

maskrcnn

. This is largely because they combine the power of deep learning with properties such as spatial invariance which is inherent to visual patterns. The generation of images using these representations demonstrates the existence of these human-like

abstract concepts in the latent space of artificial neural networks and is perhaps the only way in which non-tech experts can interface with this controversial fact. Our method of combining original but hidden artwork, subjective human input, and neural style transfer helps to broaden an insight into an artist’s creative process. Furthermore, it creates a human-AI collaboration which cultivates empathy with the creative potential of AI and its harmonious use as an artistic tool.

4 Ethical considerations

We believe a lot of trepidation surrounding machine learning is that it replaces people. We argue that the use of machine learning as an artistic tool can ultimately broaden creative insight and widen the landscape of inspirational ingenuity by human-AI collaboration. We believe that this concept is generalisable to many domains and that it implies that the jobs of tomorrow have the opportunity to be better and more fulfilling. We believe that AI art may pioneer this positive change of mentality.

References

  • [1] Pablo Picasso. The old guitarist. The Art Institute of Chicago, 1903-1904.
  • [2] Pablo Picasso. La vie. Cleveland Museum of Art, 1903.
  • [3] Mary Mathews Gedo and Picasso. A youthful genius confronts his destiny: Picasso’s "old guitarist" in the art institute of chicago. Art Institute of Chicago Museum Studies, 12(2):153–165, 1986.
  • [4] Leon A. Gatys, Alexander S. Ecker, and Matthias Bethge. A neural algorithm of artistic style, 2015.
  • [5] Pablo Picasso. La miséreuse accroupie. Art Gallery of Ontario, 1902.
  • [6] Santiago Rusiñol. Terraced garden in mallorca. Art Gallery of Ontario, 1904.
  • [7] Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross B. Girshick. Mask R-CNN. CoRR, abs/1703.06870, 2017.

Appendix

Appendix I:


Figure 3: Parc del laberint d’horta; a lost Rusiñol, reconstructed using neural style transfer.

Appendix II:


Figure 4: La femme perdue; a lost Picasso, reconstructed using neural style transfer.