YEREVAN — Robohood Inc. is a company that developed robotic painting technology. The goal of the project is to combine the capabilities of a robot and a human, shortening the path between the idea and its implementation.
In general, the process looks like this: based on the original image — a photo or an image — you can customize the style of painting, preview the visualization and start drawing in “three clicks.” Then the robot independently paints a picture on canvas with brush and oil or acrylic paints.
The difference between Robohood technology and other robo-painting projects is that Robohood is realistic. In the painting you see recognizable objects defined by the original image, not only beautiful abstractions. Robohood is easy to use, as no manual processing of each image is required — everything happens in a special interface. The software is easy even for a child to use.
Finally, it is accurate. The strokes are based on the shape and texture of the image. The final picture differs from the original picture, just as an artist’s painting differs from a photograph.
Implementing this project, it is necessary to solve different tasks: to turn a bitmap image into a vector image, taking into account the chosen style of painting — neural networks and other ML techniques are used for this; to mix paints to get the right color — this process has its own characteristics, due to different paints having different properties depending on the color and manufacturer, and their properties change during operation (in this case, computer vision is used to correct and refine colors); to convert the finished coordinates of the brushstroke into the robot’s movements so that it puts the brushstroke gently, at the right angle and at the right speed — as artists do.
