# Pre-training Enables Extraordinary All-optical Image Denoising

arXiv: https://arxiv.org/abs/2605.07810 · HTML: https://mlsanity.com/paper/2605.07810/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 59 references that are in the map.
- First posted on arXiv 8 May 2026.

## Builds on (most cited first, 24 of 24)

- [Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms](https://mlsanity.com/paper/1708.07747/) (2017)
- [Pre-Trained Image Processing Transformer](https://mlsanity.com/paper/2012.00364/) (2020)
- [Robust fine-tuning of zero-shot models](https://mlsanity.com/paper/2109.01903/) (2021)
- [MedMNIST v2 -- A large-scale lightweight benchmark for 2D and 3D biomedical image classification](https://mlsanity.com/paper/2110.14795/) (2021)
- [Rethinking ImageNet Pre-training](https://mlsanity.com/paper/1811.08883/) (2018)
- [Using Pre-Training Can Improve Model Robustness and Uncertainty](https://mlsanity.com/paper/1901.09960/) (2019)
- [Rethinking Pre-training and Self-training](https://mlsanity.com/paper/2006.06882/) (2020)
- [A Neural Representation of Sketch Drawings](https://mlsanity.com/paper/1704.03477/) (2017)
- [Sophia: A Scalable Stochastic Second-order Optimizer for Language Model Pre-training](https://mlsanity.com/paper/2305.14342/) (2023)
- [All-Optical Machine Learning Using Diffractive Deep Neural Networks](https://mlsanity.com/paper/1804.08711/) (2018)
- [Scaling Laws and Compute-Optimal Training Beyond Fixed Training Durations](https://mlsanity.com/paper/2405.18392/) (2024)
- [Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systems](https://mlsanity.com/paper/2104.13386/) (2021)
- [CelebA-Spoof: Large-Scale Face Anti-Spoofing Dataset with Rich Annotations](https://mlsanity.com/paper/2007.12342/) (2020)
- [Large-scale neuromorphic optoelectronic computing with a reconfigurable diffractive processing unit](https://mlsanity.com/paper/2008.11659/) (2020)
- [Training of Physical Neural Networks](https://mlsanity.com/paper/2406.03372/) (2024)
- [YOLO5Face: Why Reinventing a Face Detector](https://mlsanity.com/paper/2105.12931/) (2021)
- [Backpropagation-free Training of Deep Physical Neural Networks](https://mlsanity.com/paper/2304.11042/) (2023)
- [Nonlinear Processing with Linear Optics](https://mlsanity.com/paper/2307.08533/) (2023)
- [Computational Imaging Without a Computer: Seeing Through Random Diffusers at the Speed of Light](https://mlsanity.com/paper/2107.06586/) (2021)
- [YOLO-FaceV2: A Scale and Occlusion Aware Face Detector](https://mlsanity.com/paper/2208.02019/) (2022)
- [A Deep Learning Approach to Denoise Optical Coherence Tomography Images of the Optic Nerve Head](https://mlsanity.com/paper/1809.10589/) (2018)
- [All-optical image denoising using a diffractive visual processor](https://mlsanity.com/paper/2309.09215/) (2023)
- [Deeper detection limits in astronomical imaging using self-supervised spatiotemporal denoising](https://mlsanity.com/paper/2602.17205/) (2026)
- [Subwavelength Imaging using a Solid-Immersion Diffractive Optical Processor](https://mlsanity.com/paper/2401.08923/) (2024)

## Abstract (arXiv)

Optical neural networks are emerging as powerful machine learning and information processing tools because of their potential advantages in speed and energy efficiency. The training methods of these physical models, however, remain underexplored compared to their digital counterparts and are leading to suboptimal performance. This paper reports a pre-training-driven approach that leads to snapshot image denoising with substantially improved quality. We demonstrated effective free-space optical denoising by a diffractive network optimized by a two-step process including (1) pre-training using a massive dataset of 3.45 million diverse but simple images and (2) fine-tuning with the corresponding task-specific datasets. Compared to conventional Fourier-domain filtering and directly trained diffractive networks, such a transfer learning process exhibited prominent advantages for denoising images degraded by severe noise, peak signal-to-noise ratio (PSNR) below 8 dB, while preserving fine image features and improving the PSNR to above 18 dB. Importantly, the same pre-trained optical network could be consistently fine-tuned to process degraded images from highly diverse styles ranging from handwritten digits (MNIST) and chest X-rays (ChestMNIST) to CIFAR-10 images and human faces (CelebA). We further demonstrated the critical role of our optical denoisers in vision-based applications, including face detection, plate recognition, and localization of UAVs in noisy conditions.