# Joint Demosaicking and Denoising by Fine-Tuning of Bursts of Raw Images

## TL;DR

This paper introduces a method to jointly perform demosaicking and denoising directly from RAW images, enabling training on real data without ground truth RGB, and shows that fine-tuning on specific bursts enhances restoration quality.

## Contribution

It presents a novel approach to learn joint demosaicking and denoising from mosaicked RAW images without ground truth, and demonstrates the benefits of burst-specific fine-tuning.

## Key findings

- Learning from mosaicked images without ground truth is feasible.
- Fine-tuning on specific bursts improves image restoration quality.
- The method enables training with real RAW data instead of simulated data.

## Abstract

Demosaicking and denoising are the first steps of any camera image processing pipeline and are key for obtaining high quality RGB images. A promising current research trend aims at solving these two problems jointly using convolutional neural networks. Due to the unavailability of ground truth data these networks cannot be currently trained using real RAW images. Instead, they resort to simulated data. In this paper we present a method to learn demosaicking directly from mosaicked images, without requiring ground truth RGB data. We apply this to learn joint demosaicking and denoising only from RAW images, thus enabling the use of real data. In addition we show that for this application fine-tuning a network to a specific burst improves the quality of restoration for both demosaicking and denoising.