BSRAW: Improving Blind RAW Image Super-Resolution

TL;DR

This paper introduces BSRAW, a method for blind super-resolution directly on RAW images, using a realistic degradation pipeline and a new dataset, to improve real-world RAW image quality.

Contribution

It presents a novel RAW domain super-resolution approach with a realistic degradation model and a new benchmark dataset for training and evaluation.

Findings

BSRAW effectively upscales real RAW images.

The degradation pipeline improves training realism.

Results outperform sRGB-based super-resolution methods.

Abstract

In smartphones and compact cameras, the Image Signal Processor (ISP) transforms the RAW sensor image into a human-readable sRGB image. Most popular super-resolution methods depart from a sRGB image and upscale it further, improving its quality. However, modeling the degradations in the sRGB domain is complicated because of the non-linear ISP transformations. Despite this known issue, only a few methods work directly with RAW images and tackle real-world sensor degradations. We tackle blind image super-resolution in the RAW domain. We design a realistic degradation pipeline tailored specifically for training models with raw sensor data. Our approach considers sensor noise, defocus, exposure, and other common issues. Our BSRAW models trained with our pipeline can upscale real-scene RAW images and improve their quality. As part of this effort, we also present a new DSLM dataset and benchmark for this task.