As the market for artificial intelligence-related services, major tech companies began to invest in this area. An artificial intelligence-related project which was published as a demonstration developed by Meta by 2021 is now open source in a new decision.
You’ll be able turn images into motion.
The animation project was created through Meta’s Fundamental AI Research (FAIR) team, and whose demonstration was released in 2021 was released as open source following Meta’s decision. The demo on the web that was that was released in 2021 challenges users to draw a figure inside a tiny box and then Artificial Intelligence model transforms your drawings into animated sequences of walking, dancing, laughing and leaping.
The project, which is referred to as Animated Drawings, uses object detection models as well as pose estimation models and segmentation techniques based on image processing to produce a digital copy of the drawing. Computer graphics methods are used to transform and then animate the image. Meta has further refined the model by using 1.6 million drawings from users during the demonstration which suggests that the project is now going to become more robust. Additionally, during the demonstration Meta stated that only human figures would be used, but it appears that the project has also been working on a variety of designs, such as logos, anime characters, and fish.
In an article on their blog, Meta talked about the advantages of open source communities and stated, “By publishing models and code as open source, it provides a starting point for project developers to develop and expand the project, to foster a culture of innovation and collaboration in the open source community.”