Google unveils Lumiere generative AI to create more realistic images and videos from text

[Redaktion]

Google has unveiled Lumiere - the latest in realistic text-to-image and text-to-video generation using machine learning. A key innovation is the ability to create realistic motion such as walking that current generative AIs have trouble with. The software does this by creating all video frames at once rather than using keyframes and training to learn how moving objects should appear.

https://www.notebookcheck.net/Google-unveils-Lumiere-generative-AI-to-create-more-realistic-images-and-videos-from-text.798878.0.html