Just a few months ago, we talked with excitement and amazement about these AIs that could generate images from texts. Today it seems that technology has taken leaps and bounds, and the fact is that … we already have models that can create videos based on the same principle, so Meta presents us with its Make-A-Video artificial intelligence, and the videos it creates are so as amazing as it is frightening.
As with other similar templates, Make-A-Video prompts you to enter a description of what you want to create. So, typing “A dog wearing a red superhero cape and flying in the sky” will give you the expected result. Keep in mind that this technology is still in its infancy and the generated videos can be interesting to say the least.
Make-A-Video is not yet available to the public. However, some have already tried it. Although new, the results are impressive and we can’t wait to see how this artificial intelligence will evolve over the years. Like AI creating images, they could soon replace some of the Internet’s most popular areas, such as image and video banks.
Hey Make-A-Video, I want you to draw a couple in the rain
Meta has managed to develop a powerful tool. However, for this artificial intelligence to work, it is necessary to use very powerful computers. Don’t forget that these image-generating AIs already required a lot of technical resources; now a device capable of turning text into video requires much more.
Why so much power? Remember that a video is just a collection of images glued together with audio embedded. Now imagine how long it takes the AI to create one frame, and multiply that by the number of frames per minute of the video (there could be several thousand). Add to that the fact that all of these generated images must be combined into one file. This is definitely insane.
According to Tanmai Gupta, a computer vision researcher at the Allen Institute for Artificial Intelligence, the results obtained with Meta Make-A-Video AI are very promising. In addition, it demonstrates the ability of the model to capture 3D objects. As the camera rotates, new details of the subject and background appear. It also demonstrates that the AI is able to distinguish between depth and light sources.
However, Gupta adds that the research community still has a long way to go, especially if these systems are used for professional video editing and content creation. He also adds that the technology is still trying to create an interaction between objects in a scene.
The Make-A-Video research is based on the latest advances in text-to-image technology designed to enable text-to-video conversion. The system uses images with descriptions to find out what the world looks like and how it is usually described.
It also uses unmarked videos to see how the world moves. Using this data, Make-A-Video allows you to bring your imagination to life, creating whimsical and unique videos from just a few words or lines of text.
One of the most amazing aspects of this artificial intelligence is its ability to create without the need for paired text and video data. So far, many image generators have been based on content galleries that already combine text and video. Make-A-Video, however, does not require so much information to work, which demonstrates a significant advantage.
This AI can be used in many different ways. Whether it’s giving movement to a single image or filling a sequence of images with movement. In addition, you can also create variations of the video from the original. The style you require, such as DALL-E or Midjourney, is up to your imagination.