Paper notes-Image Denoising via CNNs: An Adversarial Approach-优快云博客

本文介绍了一种利用卷积神经网络(CNN)进行盲图像去噪的方法。该方法提出了一种新的多尺度自适应CNN架构，能够从噪声图像中提取足够特征并重建清晰图像，其性能与当前最先进的去噪技术相当。实验表明，结合干净和噪声图像进行训练能提高特征映射的质量。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

Paper notes-Image Denoising via CNNs: An Adversarial Approach

1.Main task

Is it possible to recover an image from its noisy version using convolutional neural networks?The paper propose a new CNN architecture for blind image denoising which synergically combines three architecture compoents,a multi-scale feature extraction layer which helps in reducing the effect of noise feature maps.

Image denoising is a fundamental image processing problem whose objective is to remove the noise while preserving the original image structure.Traditional denoising algorithms are given some information about the noise,but the problem of blind image denoising involves computing the denoised image from noisy one with out any knowledge of the noise.

This paper addresses how CNNs can be used for blind image denoising.They are noisy images and require the network to gather enought features from this images so that a noise-free version can be computed from them.

1,They propose a multi-scale adaptive CNN architecture which gives a competitive performance to the state-of-the-art image denoising approaches.

2,A train regime which exploits clean images as well as noisy images to get good feature maps for recongstruction

2,background

3,Main solution