For a few years now, generating impressive artworks have been one of the many popular applications in the world of AI. Also, Generative Adversarial Networks (GAN) have been popularly applied by researchers and artists for artistic tasks, including creating artworks, style transfers, converting photographs into beautiful painting and more.
In one of the popular instances, in 2018, a generative adversarial network (or GAN) generated portrait entitled “Edmond De Belamy” was sold for $432,500, which was 40 times the initial expectation. Before we dive into how AI is making a huge difference in the world of art, let’s look at some of the latest AI-based art innovations that have made its way in 2020.
Sketch-And-Paint GAN (SAPGAN)
Recently, undergrad student Alice Xue from Princeton University proposed a GAN-based model — SAPGAN — that generates Chinese landscape paintings from end to end, without any conditional input. SAPGAN is composed of two GANs, SketchGAN for the generation of edge maps, and PaintGAN for subsequent edge-to-painting translation. According to the researchers, a 242-person Visual Turing Test study reveals that SAPGAN paintings are mistaken as human artwork with 55% frequency, significantly outperforming paintings from baseline GANs.
Know more here.
GANKsy Street Art
The creators of GANKsy trained a StyleGAN2 neural network by using the portfolio of renowned street artist Banksy. GANKsy was created in September this year and underwent a strenuous AI training regime using hundreds of street art photos for thousands of iterations to become the fully-formed artist. Currently, 256 masterpieces are for sale, starting at 1 pound.
Know more here.
AI Gahaku
AI Gahaku is an online AI artist tool that has the capability to transform a normal photograph into renaissance painting. In March this year, a Tokyo-based full-stack developer named Sato introduced this AI tool via micro-blogging platform Twitter.
Know more here.
AI Art Studio: Platform
Introduced in May this year, Playform is an AI Art studio that allows artists to experiment as well as explore the use of generative AI as part of their creative process. According to sources, the goal of its developer is to make AI accessible to artists, realising several challenges that are being faced by artists and creatives when approaching this technology.
Know more here.
AI Tool That Recreates Art
In the month of June, researchers from MIT introduced a new AI system that can synthesise time-lapse videos of paintings, known as Timecraft. It is a neural network model that is being implemented using a sequential conditional variational autoencoder (CVAE) framework. According to sources, the model is trained using CVAE to reconstruct a real frame while building upon its previous predictions.
Know more here.
How AI Is Revolutionising Art
Looking back to last year, it is clear that AI is playing a significant role in revolutionising the world of art. Some of the groundbreaking instances from last year include AI researchers at Samsung used deepfake techniques to create living portraits of Monalisa. Also, the world’s first ultra-realistic robot artist developed by Aidan Meller. The AI robot can draw as well as paint using pen and pencil and creates sketches in a human-like manner.
Besides the innovation of AI tools, AI is also making sure of enhancing the creativity among artists as well as the art business. With its potential, the technique has been fundamentally changing the business landscape of AI.
Below here are some of the ways that AI is helping to make a better change in art in 2020:
- With the help of predictive pricing, AI is assisting artists in predicting the shipping cost of a masterpiece.
- AI is not only evaluating and predicting the value of an artwork, but also it is helping to forecast the value of the next art piece.
- With AI assistance, an artist can avoid repetitive tasks. This way, the artist will get more time to focus on various new creative ideas.
- AI tools are data-hungry. The more you feed data, the better outcome you will receive. This means if an artist feeds its AI tool with lots of artwork from various famous artists, the AI will imitate the patterns of the various artworks and in return will produce unimaginable and unique artwork.
- Not only AI has found its way into paintings, but also it has reached a point where it is creating music from scratch. With the help of reinforcement learning, neural networks, among others, AI is creating lyrics, music as well as recommending music to listeners.
