ByteDance launches OmniHuman AI to create realistic videos from a single photo

ByteDance has introduced OmniHuman, an AI that creates realistic videos from a single photo. This new system can animate full-body movements and gestures, improving on previous models that only animated faces or upper bodies. OmniHuman was trained on over 18,700 hours of human motion data, using a method that combines text, audio, and body movements. This approach allows for more diverse and effective training compared to earlier techniques. The technology could impact digital entertainment and communications significantly. However, it also raises concerns about the potential for misuse in creating misleading synthetic media.