Google’s New AI Photo Upscaling Tech

Google researchers in the company’s Brain Team highlight new accomplishments in image super-resolution in an article titled “High Fidelity Image Generation Using Diffusion Models” published on the Google AI Blog (and spotted by DPR).

A machine learning model is taught to convert a low-resolution photo into a detailed high-resolution photo in image super-resolution, and potential applications range from recovering old family images to improving medical imaging.

The approach is called SR3, or Super-Resolution via Repeated Refinement.

“SR3 is a super-resolution diffusion model that takes as input a low-resolution image, and builds a corresponding high resolution image from pure noise. The model is trained on an image corruption process in which noise is progressively added to a high-resolution image until only pure noise remains.”

“It then learns to reverse this process, beginning from pure noise and progressively removing noise to reach a target distribution through the guidance of the input low-resolution image.” Google added.

photo by:

The new technology is incredible, and Google is proud of its team for doing such a fantastic job. We can’t wait to see what else they can do in the world of diffusion models.

Leave a Comment

Your email address will not be published. Required fields are marked *