Adobe Reveals AI Upscaler That Can Improve Video Resolution By 8x

Adobe has unveiled a groundbreaking AI innovation poised to revolutionize video resolution enhancement. Dubbed VideoGigaGAN, this new generative AI upscaler boasts the remarkable capability of elevating low-resolution and blurry videos by up to eightfold. In a recent publication by Adobe on April 18, the prowess of VideoGigaGAN is showcased, demonstrating remarkable results even with poor source imagery.

Notably, Adobe asserts VideoGigaGAN's superiority over existing video super-resolution methods, attributing its success to the avoidance of common AI artifacts and the generation of finely detailed outputs. Leveraging the foundation of the expansive upscaler GigaGAN, VideoGigaGAN further enhances video quality by introducing temporal modules and refining techniques to ensure temporal consistency in unsampled videos.

As highlighted by The Verge, while Generative Adversarial Networks (GANs) have proven effective in upscaling still images, their application to videos has historically resulted in unwanted flickering. Despite alternative upscaling methods, none have managed to achieve the finesse exhibited by GANs.

Adobe's demonstration of VideoGigaGAN's capabilities, where it transformed 124×124 videos into stunning 1024×1024 resolutions, underscores the AI's transformative potential, rendering it challenging to discern the intervention of AI wizardry. At present, VideoGigaGAN remains a research preview without immediate plans for integration into Premiere Pro or the broader Creative Cloud suite. This revelation follows closely on the heels of Microsoft's announcement of an AI super resolution, "Automatic Super Resolution," slated for integration into Windows 11.

In the realm of Adobe and AI, this unveiling marks yet another stride in the company's pursuit of innovative solutions. Earlier this week, Adobe introduced one of its most significant updates to Photoshop, anchored in the advancements achieved through the Adobe Firefly Image 3 Foundation Model. The beta release of the third iteration of Firefly promises substantial enhancements in quality and control, empowering users with higher-quality image generation, improved prompt comprehension, and enriched detail and variety.

Ashley Still, senior president of the Creative Product Group at Adobe, lauds the transformative potential of this update, emphasizing its capacity to elevate creativity to unprecedented heights. Through the integration of Firefly-powered features into Photoshop, Adobe aims to bridge the gap between ideation and creation, making image manipulation more potent and accessible to all.

To learn more about this groundbreaking development, visit the source article on PetaPixel: https://petapixel.com/2024/04/25/adobe-reveal-ai-upscaler-that-can-improve-video-resolution-by-8x/.

Additionally, for high-end digital marketing, app creation, AI-powered solutions, chatbots, software development, and website creation services, explore the innovative offerings of Spectro Agency at spectroagency.com