A recent patent filing by Nintendo indicates the company is developing an AI upscaling technology akin to Nvidia’s DLSS, possibly in preparation for the Nintendo Switch 2.
The technology, described in a U.S. patent published this past Tuesday, enhances images using a trained neural network. It suggests significant improvements in-game image quality for the upcoming console.
The patent details a method for dividing a source image into blocks and adding contextual data. These blocks are then processed through various channels within an activation matrix and run through a trained neural network to produce a higher-resolution image.
The approach mirrors Nvidia’s DLSS (Deep Learning Super Sampling) technology, which uses AI to upscale lower-resolution images to higher resolutions without traditional computational costs. Thus, it improves image quality and performance in video games. AMD employs similar technologies with its FSR and Intel with XeSS, enhancing performance on less capable hardware.
AI upscaling is also extending to gaming consoles. For instance, the newly introduced PlayStation 5 Pro features the PlayStation Spectral Super Resolution (PSSR), which significantly applies machine learning to refine image details.
The timing of Nintendo’s patent publication aligns with reports suggesting the Nintendo Switch 2 will incorporate DLSS-style technology to boost its performance capabilities. The launch of the Nintendo Switch 2 is anticipated before the end of March, with Nintendo confirming that the new hybrid console will be backwards compatible with existing Switch games. Recent leaks have pinpointed the release date as March 28.
The development marks a significant step for Nintendo as it prepares to introduce next-generation technology to its console lineup, enhancing gaming experiences with advanced image upscaling capabilities.
