# Rethinking Learning-based Demosaicing, Denoising, and Super-Resolution Pipeline

arXiv: https://arxiv.org/abs/1905.02538 · HTML: https://mlsanity.com/paper/1905.02538/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 May 2019.

## Built on by (most cited first, 13 of 13)

- [PIDSR: Complementary Polarized Image Demosaicing and Super-Resolution](https://mlsanity.com/paper/2504.07758/) (2025)
- [Pix4Point: Image Pretrained Standard Transformers for 3D Point Cloud Understanding](https://mlsanity.com/paper/2208.12259/) (2022)
- [Unsupervised Spectral Demosaicing with Lightweight Spectral Attention Networks](https://mlsanity.com/paper/2307.01990/) (2023)
- [Efficient HDR Reconstruction from Real-World Raw Images](https://mlsanity.com/paper/2306.10311/) (2023)
- [Joint Demosaicing and Denoising with Double Deep Image Priors](https://mlsanity.com/paper/2309.09426/) (2023)
- [Event Camera Demosaicing via Swin Transformer and Pixel-focus Loss](https://mlsanity.com/paper/2404.02731/) (2024)
- [DSDNet: Raw Domain Demoir'eing via Dual Color-Space Synergy](https://mlsanity.com/paper/2504.15756/) (2025)
- [Diff-Unfolding: A Model-Based Score Learning Framework for Inverse Problems](https://mlsanity.com/paper/2505.11393/) (2025)
- [Multispectral Demosaicing via Dual Cameras](https://mlsanity.com/paper/2503.22026/) (2025)
- [Why Low-Light Cameras Go Color Blind: Removing Color Bias in Raw Denoising](https://mlsanity.com/paper/2607.11090/) (2026)
- [StereoISP: Rethinking Image Signal Processing for Dual Camera Systems](https://mlsanity.com/paper/2211.07390/) (2022)
- [SDAT: Sub-Dataset Alternation Training for Improved Image Demosaicing](https://mlsanity.com/paper/2303.15792/) (2023)
- [Efficient Bayer-Domain Video Computer Vision with Fast Motion Estimation and Learned Perception Residual](https://mlsanity.com/paper/2508.05990/) (2025)

## Abstract (arXiv)

Imaging is usually a mixture problem of incomplete color sampling, noise degradation, and limited resolution. This mixture problem is typically solved by a sequential solution that applies demosaicing (DM), denoising (DN), and super-resolution (SR) sequentially in a fixed and predefined pipeline (execution order of tasks), DM→DN→SR. The most recent work on image processing focuses on developing more sophisticated architectures to achieve higher image quality. Little attention has been paid to the design of the pipeline, and it is still not clear how significant the pipeline is to image quality. In this work, we comprehensively study the effects of pipelines on the mixture problem of learning-based DN, DM, and SR, in both sequential and joint solutions. On the one hand, in sequential solutions, we find that the pipeline has a non-trivial effect on the resulted image quality. Our suggested pipeline DN→SR→DM yields consistently better performance than other sequential pipelines in various experimental settings and benchmarks. On the other hand, in joint solutions, we propose an end-to-end Trinity Pixel Enhancement NETwork (TENet) that achieves state-of-the-art performance for the mixture problem. We further present a novel and simple method that can integrate a certain pipeline into a given end-to-end network by providing intermediate supervision using a detachable head. Extensive experiments show that an end-to-end network with the proposed pipeline can attain only a consistent but insignificant improvement. Our work indicates that the investigation of pipelines is applicable in sequential solutions, but is not very necessary in end-to-end networks. RRCode, models, and our contributed PixelShift200 dataset are available at urlhttps://github.com/guochengqian/TENet