基于自编码器理论实现的图像去噪方法研究

Research on Image Denoising Method Based on Auto-encoder Theory in Machine Learning

为研究一种新型图像去噪的手段,采用机器学习中的自编码器模型对图像进行优化。通过对原始的清晰图像添加噪声,作为编码器模型的输入数据,未添加噪声的原始图像作为标准值,对编码器模型进行训练以提高其图像去噪能力。结果表明,经过一定程度训练的自编码器模型,具备了一定的图像去噪功能。

In order to study a new method of image denoising,the auto-encoder module in machine learning is used to optimize the image.By adding noise to the original clarity image,being taken as the input data of the auto-encoder module,and the original image without adding noise as a standard value,the auto-encoder module is trained to improve image denoising ability.Experimental results show that the auto-encoder module has a certain image denoising function after training.