Denoising: from classical methods to deep CNNs

TL;DR

This paper reviews the evolution of image denoising from classical techniques to modern deep CNNs, emphasizing the role of neural networks like U-Net and score diffusion in improving denoising performance.

Contribution

It provides a pedagogical overview of denoising methods, highlighting the transition from classical approaches to deep learning and the importance of score diffusion in image generation.

Findings

Deep CNNs like U-Net outperform classical methods in denoising tasks.

Score diffusion is crucial for effective image generation and density estimation.

Neural networks adapt well to various image types, achieving optimal results.

Abstract

This paper aims to explore the evolution of image denoising in a pedagological way. We briefly review classical methods such as Fourier analysis and wavelet bases, highlighting the challenges they faced until the emergence of neural networks, notably the U-Net, in the 2010s. The remarkable performance of these networks has been demonstrated in studies such as Kadkhodaie et al. (2024). They exhibit adaptability to various image types, including those with fixed regularity, facial images, and bedroom scenes, achieving optimal results and biased towards geometry-adaptive harmonic basis. The introduction of score diffusion has played a crucial role in image generation. In this context, denoising becomes essential as it facilitates the estimation of probability density scores. We discuss the prerequisites for genuine learning of probability densities, offering insights that extend from mathematical research to the implications of universal structures.