Rain, smoke, dirt, and debris-produced distortion would normally be destructive for any videographer's footage, but global research and development firm Cambridge Consultants is now offering an AI solution to reconstruct damaged or obscured frames in real-time. Dubbed DeepRay, this new program calls to mind Adobe’s distortion-correcting system, but manages to accomplish the same feat but in real-time and with live video.
Cambridge Consultants currently plans on showcasing DeepRay at CES 2019. However, in its current state, the system takes advantage of a machine learning architecture known as generative adversarial network (GAN). Speaking broadly, GAN is a two-part neural network that consists of generators and discriminators. DeepRay features a total of six networks, each with their own team of generators and discriminators, which each analyze footage to remedy damaged frames,
Key Themes Behind This Trend
- Real-time Video Restoration
- Cambridge Consultant's new AI program DeepRay is able to reconstruct damaged or obscured frames in real-time, presenting opportunities for filmmakers and video production companies to deliver high-quality content even in challenging filming environments.
- Machine Learning for Video Enhancement
- The use of generative adversarial network in DeepRay presents opportunities for businesses in the field of AI and machine learning to innovate and create similar solutions for other industries that may require real-time video restoration, such as security and surveillance.
- Live Video Distortion Correction
- DeepRay's ability to correct distortions in live video could revolutionize fields such as live broadcasting and streaming, presenting opportunities for businesses to provide seamless and high-quality real-time broadcasts or streams.
Where This Applies
- Filmmaking and Video Production
- DeepRay's real-time video restoration capability could revolutionize the way filmmakers and video production companies work, as it enables them to deliver high-quality content even in challenging filming environments.
- Artificial Intelligence and Machine Learning
- The use of generative adversarial network in DeepRay presents opportunities for businesses to create similar solutions for other industries that require real-time video restoration, such as security and surveillance.
- Live Broadcasting and Streaming
- DeepRay's ability to correct distortions in live video presents opportunities for businesses in the field of live broadcasting and streaming to provide seamless and high-quality real-time broadcasts and streams.
