神经网络在建立压缩图像退化模型中的应用

Application of artificial neural network in building degraded model of compressed image

在有损压缩过程中,为了得到较高的压缩比,一般牺牲了图像的质量,为了得到高的压缩比,同时图像质量也不会过分下降,本文主要利用人工神经网络来估计压缩图像过程导致的图像信息损失过程,我们称之为退化模型。在估计了压缩图像的退化模型后,对应的可以估计其重建模型。在此基础上,利用图像超分辨率重建的思想来部分恢复图像的质量。得到了图像的退化模型和重建模型后,我们可以利用这些模型来改善图像的质量。实验证明,利用人工神经为了估计图像退化模型比较有效,压缩图像的质量可以得到较大提高。

During the compression of image, to get a high compressed ratio, people always only get a degraded image. This article aims to build the degraded model of the compressed image through using the artificial neural network. After obtaining the degraded model, we can also build the constructed model of the compressed image. Then, we can improve the quality of the compressed image on the basis of image super-resolution technology. The experiments show that this method is effective and by this method we can improve the quality of compressed image.