Chimera Painter by Google AI can turn your doodles into realistic monsters
Chimera Painter by Google AI is a demo tool that uses machine learning (ML) to allow users to create a developed creature based on a rough sketch. Chimera Painter adds features and textures to the outline of a creature, segmented with body part labels such as claws or horns. The ML model is built with the Unreal Engine and the ML algorithm based on GAN (generative confrontational network). Although the tool is primarily aimed at game developers, anyone can use it to turn a rough drawing into a realistic looking monster, just for fun.
What if artists had a brush that acted less like a tool and more like an assistant? A machine learning model that acts like a paintbrush could reduce the amount of time required to create high-quality art without sacrificing artistic options, perhaps even enhancing creativity. ” Google The AI team wrote in a blog post, explaining the Chimera Painter demo in detail.
Chimera painter You have four preset creature outlines to choose from. These can be modified using the tools available on the left of the screen, of 20 different parts of the body such as head, horns, teeth, claws, wings, tails and more. Each one has a specific color assigned to it that will help you differentiate between body parts when drawing. You can adjust the brush size and use the eraser and transparent tools as well.
Once you are done creating, click Transform to turn your doodle into a realistic looking creature. You can also use the tools to add any part anywhere; you can add horns on your monster’s torso and teeth on its tail, and the tool will do its best to incorporate that.
You can also load a creature outline created in an external program such as Photoshop. Download one of the preset creature outlines to get the required colors needed for each part of the creature and use it as a template. Then click the Load in the demo to use that scheme.
Google said that to pursue this goal, the team created a fantasy digital card game prototype around the concept of combining creatures into new hybrids that can then battle each other. To train adversarial generative networks (GAN), they created a dataset of full-color imagery with outlines of single-species creatures that were adapted from 3D creature models. After training the model, it was tasked with generating multi-species chimeras that were based on the contours provided by the artists.
The best-performing model was later brought into Chimera Painter, Google’s AI team explained. A new artist-led semi-automated approach was also developed to create an ML training dataset from 3D creature models.
Google is not the first to create a tool of this type. NVIDIA has a similar deep learning mode that turns squiggles into realistic landscapes, and IBM use GAN paint to do the same for buildings.