Submit a Project

ML x ART

A showcase with creative machine learning experiments

👋 I'm Emil, a resident at the Google Arts & Culture Lab. I've curated this collection.

Year
Artist
Technology
Theme
Medium
Default order

Travis LeRoy Southworth

Using the artist's own artworks to create a Style Transfer on portraits found online and animated in After Effects

Style Transfer

Sam Lavigne

Using semantic segmentation trained on Figaro-1k to extract all the haircuts of Mark Zuckerberg

Semantic segmentation, Figaro-1k

Mario Klingemann

A next frame prediction using pix2pix feeding the newly generated image repeatedly

pix2pix, GAN

Yotam Mann

A WaveGAN-inspired model trained on the acapella vocal ensemble in Košice, Slovakia, translating Glenn Gould's piano into a Gregorian Choir

WaveGAN, GAN

Obvious

GAN trained on masks, and make the masks according to the GAN-generated images

GAN

Kevin Ho, Google Creative Lab

Measure the distance between CNN features to pair fonts

RNN, NNS

Hector Rodriguez

Reconstruct the film Alphaville (1965) by extracting patches from the same film and reassemble them with Simplex Volume Maximization

Simplex Volume Maximization

Guillaume Slizewicz

A character generator model is trained on a technical database to identify criminals from 1907. It's contrasted with a modern webcam linked to a facial recognition algorithm

Face recognition

Google ZOO, bit.studio

Learn shadow art by having a CNN classify and correct your hand patterns

CNN, Classification, Camera

Jonathan Fly

Use SinGAN to train a GAN with a single image, Frymire, the classic compression image

SinGAN, GAN

Helena Sarin

Apply CycleGAN to a Sketch of the Fonthill Castle

CycleGAN, GAN

David Ha, Jonas Jongejan, Ian Johnson, Google Creative Lab

An RNN model generates doodles based on your style

RNN

Christine Sugrue, Google Arts & Culture Lab

Use Object Detection to explore details in artworks

CNN, Object Detection

François Quévillon

A tool for creating an endless car video. It's made by modifying parameters in the sampling system that's created with image features and other technologies

CNN

Aven Le Zhou

A pix2pix model that translates drawings to Chinese “Shanshui” artworks

pix2pix

Vibert Thio, Douglas Eck

An interactive room where you can create music with Magenta's javascript version of MusicVAE and MelodyRNN

Magenta, VAE, RNN

Yotam Mann, Google Creative Lab

A user plays a few keys on a piano and an RNN model generates more according to your input

RNN, Magenta

Ardavan Bidgoli, Pedro Veloso

VAE-generated 3D point clouds for different objects, interpolated with a MIDI controller

VAE, Hardware, Point Cloud

Albert Jimenez, Adria Romero-Lopez, Mohamed Akrout, Stephen Solis-Reyes, Anirudh Challa

Use CycleGAN to turn Mars' surface into earth-like

CycleGAN, GAN

Bastien Girschig, Romain Cazier, Google Arts & Culture Lab

Match a user-generated shape with paintings; the CNN features of the shapes are matched with a database of features of outlined artworks

NNS, Edge Detection, CNN

Vadim Epstein

CLIP with a custom FFT-based generator turning phrases into paintings in the format of a meditative story from famous art titles

CLIP, FFT

Jason Antic

Colorize the B&W movie Metropolis with a pix2pix U-Net

U-net, Resnet, ImageNet, pix2pix, DeOldify

Mike Tyka

Create videos with BigGAN and add music with the Music Transformer

BigGAN, GAN, Music Transformer

Vadim Epstein

Painting by Viktor Vasnetsov, animated with 'Animating Landscape' and Multi-Domain Multi-Modality I2I translation

Multi-Domain Multi-Modality I2I translation, Animating Landscape

Submit projects here 💅

☞ ML x ART on Twitter

Copyright © 2021 Emil Wallner

All artworks © of their respective owners