The new techniques use machine learning (ML) and artificial intelligence (AI) algorithms that go well beyond simple algorithms and nearest neighbor scaling concepts. ML techniques use computers to classify types of images and compare low- and high-resolution versions of them to develop tools to allow the computer to then reconstruct a high-fidelity image from a low-fidelity one. AI or “Deep Learning” takes this a step further by adding a feedback loop allowing the system to learn which reconstructions were better than others. The big benefit: the algorithms can get better over time.
Upscaling and image restoration is a very important feature for an 8K TV as it allows images with sharper edges, more texture, reduced jaggies and reduced noise. Many of the providers of 8K TV will implement some form of this image restoration technology, but not all implementations will be equal. Samsung has been showing their image restoration capabilities for some time and they highlight the ability to improve 2K and 4K images to 8K fidelity while also reducing image noise and eliminating compression artifacts.
About the author: