The point where the talent of artificial intelligence has reached today and continues to surprise us with each passing day. This time, the new news came from MIT, and the researchers managed to create videos that show how artists use their hands to paint.
Experts doing research at MIT aimed to bring together artificial intelligence and art this time and succeeded in these goals. According to the incoming information, the team developed an artificial intelligence that estimated the artists’ works using watercolors and how they make digitally generated images, and how they use their hands, and put it on video. It was reported that the developed artificial intelligence was trained with artistic videos available on Vimeo and YouTube sites.
It was said that the probable model that emerged was able to perceive the steps in the construction process by looking at a single image of a work of art. When we think that most works recorded in the field of painting were made long before today, it is possible to say that this is a very exciting development. Seeing the works of art through the eyes of real artists and watching the stages of their construction are extremely amazing.
This artificial intelligence is a little different from others:
It is possible to convey that the researchers who carried out the study agreed with art lovers who heard the news. According to the researchers’ statement, this tutorial research, which makes us feel like we are next to artists who paint using unique brush strokes and color combinations, has interesting elements for everyone. At the same time, it was added to the information given that this artificial intelligence developed is different from other artificial intelligence forms. The difference is explained by the fact that other artificial intelligence tend to focus on shorter human movements.
The study is an average of four minutes long and contains 117 digital images. The number of watercolor paintings is 116 and takes 20 minutes in total. The whole study; landscape and still life concentrates on works. In the analysis, it seems that the work is extremely successful in transferring the works of real artists. Researchers still continue their efforts to improve the scope of this work and to train artificial intelligence for other works.