Nearly eight years after Grand Theft Auto V was released, modders are still enriching — and ruining — the game‘s celebrated visuals. Last week, researchers from Intel Labs unveiled an AI-powered revamp of the Rockstar classic that brings the graphics close to photorealism. TNW spoke to Intel research scientist Stephen Richter to find out more about their technique — and the potential to productionize the method. The cornerstone of the Intel Labs method is a convolutional network, a deep learning architecture that’s commonly used for image processing. The team trained their convolutional networks on real-life images to translate GTA V’s…

This story continues at The Next Web

Or just read more coverage about: Intel