ByteDance, the parent company of TikTok, has introduced OmniHuman, an AI model capable of creating highly realistic videos of people performing various actions, such as talking, gesturing, and even playing instruments, using just a single photograph as input, reports Forbes.
ByteDance AI: Impressive Capability
The OmniHuman-1 project, developed by ByteDance, has demonstrated impressive capabilities in generating highly realistic human videos from limited input, particularly audio.
The model supports diverse image inputs, including portraits, half-body, and full-body images, enabling the generation of lifelike results across various scenarios.
While not yet publicly available, the research paper showcased impressive sample videos, demonstrating the model’s ability to generate realistic hand and body movements from multiple angles.
The model can also animate characters, bring historical figures to life, and create seemingly authentic footage of past events, such as a video of Albert Einstein delivering a lecture with realistic facial expressions and hand gestures.
This advancement positions ByteDance and TikTok as key players in the race to develop the most realistic AI-generated human figures.
These digital figures are increasingly prevalent, appearing in various applications, including product endorsements by virtual influencers, government services provided by AI-powered avatars, and even in the realm of entertainment with AI-generated versions of celebrities interacting with fans.
Extensive Training
The OmniHuman-1 model was trained on a massive dataset of over 18,700 hours of human video data, incorporating diverse inputs like text, audio, and physical poses. This extensive training dataset, combined with advanced AI techniques, enables the model to generate highly realistic and nuanced human movements and expressions.
While the use of such vast datasets raises concerns about potential privacy implications, ByteDance has clarified that the model was not trained on data from its own products like TikTok.
The company has also emphasized the importance of responsible AI development, stating that any future public release of OmniHuman-1 will include safeguards against the creation of harmful or misleading content. Additionally, measures will be implemented to ensure transparency and make it clear that the generated content is AI-created.
The development of OmniHuman-1 highlights the rapid advancements in AI technology and its potential to revolutionize various fields, from entertainment and education to social interaction and communication. However, it also underscores the critical need for ethical considerations and responsible development to mitigate potential risks and ensure the beneficial use of these powerful technologies.
Did you subscribe to our daily Newsletter?
It’s Free Click here to Subscribe!
Source: Forbes
