神经网络在彩色图像压缩中的应用

文中旨在提出一种基于神经网络的图像压缩算法对彩色图像信息进行处理，从而减少大规模彩色图像的冗余度，方便其传输、存储及加密等。该算法通过将BP（Back Propagation）神经网络用于彩色图像压缩，利用其多层前馈网络的模式变化能力，实现了对由RGB编码得到的彩色图像数字矩阵进行的压缩编码。经Matlab仿真实验表明，该算法具有良好的压缩效果，且与灰度编码下的图像压缩结果对比，具有更好的压缩效率及保真效果，并能有效地保留原彩色图像的色彩信息，能够满足彩色图像压缩处理的要求。

Application of Neural Network in Color Image Compression

It is aimed at proposing a new color image compression algorithm to process color image information, reducing the redundancy of color image information and making it convenient for images transmission and storage. To accomplish this goal, it applies BP neural network to color image compression. After obtaining the image matrices using RGB coding method, the algorithm achieves image com- pression coding by use of mode-transformation ability of multilayer feedforward, neural networks, which is one property of BP { Back Propagation ） neural network. And through Matlab simulation experiment, the results show that the algorithm is effective in color images compression. Moreover, a comparison with compressing the color image of grey coding method based on the same algorithm, shows that proposed algorithm has higher compression efficiency and fidelity effect, also can retain the color information of the original image in a large degree which meets the demands of processing color images.