JPEG基线算法对BP神经网络图像压缩处理的影响

根据BP神经网络图像压缩处理中,存在对图像信息高低频部分处理质量不同和边缘效应等问题,提出了采用JPEG基线算法于BP神经网络图像压缩处理结构中,建立了该系统。并采用灰阶Lena图像进行实验,通过实验分析发现,采用这种新的结构来处理图像,不仅可以得到较大的压缩比,而且具有较好的峰值信噪比。实验结果证明这种具有自适应性的图像处理方法,不仅可行,而且能高效、稳定地重建图像。

Image Compression of BP Neural Network with JPEG Baseline Algorithm

In this paper,we mainly research that during BP neural network image compression processing the high-low spatial frequencies distribution differences of an image lead to the poor edge quality problem of reconstructed image.Therefore,we present a new image processing structure of BP neural network image compression with JPEG baseline algorithm,and established this image processing system.Through making some experiments using a grayscale Lena image,the experiment results tell us that this processing structure brings not only the higher compression ratio,but also the better PSNR.Moreover,it also testifies that this adaptive image processing method is feasible,and can reconstruct image efficiently and steadily.