Adobe’s recent focus on AI-driven features continues with the introduction of VideoGigaGAN, an experimental AI technology that can upscale video footage by eight times without introducing the usual artifacts or flickering associated with this process. VideoGigaGAN surpasses existing Video Super Resolution (VSR) methods by eliminating these artifacts and enhancing sharpness and detail, which most other systems struggle to achieve simultaneously.
The success of VideoGigaGAN stems from its foundation on the GigaGAN image upsampler and the incorporation of various techniques, including temporal attention (reducing artifacts), feature propagation (adding detail), anti-aliasing, and HF shuttle (transferring high-frequency features). This combination results in upscaled video footage with impressive levels of detail, such as realistic skin textures, fine hairs, and intricate feather details.
The potential applications of VideoGigaGAN in video editing software like Premiere Pro and After Effects are significant. It could empower video producers to transform low-resolution shots into visually appealing and detailed footage. However, the ethical implications of using AI to enhance human appearances must be carefully considered.
While Adobe has not yet announced plans to integrate VideoGigaGAN into its products, it is clear that the company is investing heavily in AI-powered video enhancement. This move is part of a broader trend in the industry, with companies like NVIDIA, Microsoft, and Blackmagic Design also developing their own upscaling technologies.
As AI continues to advance, we can expect to see further innovations in video enhancement and other creative fields. VideoGigaGAN is a testament to the potential of AI to transform the video production process and open up new possibilities for filmmakers and video editors.
