摘要
作者是在已有的基于 DCT变换的双 VQ静止图像算法的基础上提出的改进算法 ,利用神经网络中的自组织特征映射神经网络进行图象矢量量化的码书设计 ,通过实验表明本算法在进一步提高压缩比的同时可以得到令人满意的图像效果。
In the paper, a modified compression algorithm is proposed based on still picture compression algorithm using two vectorquantization by DCT and Self-Organizing Feature Map neural networks are applied to design vector quantization codes. It was shown that the modified compression algorithm can enhance the compression ratio on the condition that the vectorquantization error has been reduced in the experiment, while get the satisfying rebuilt still picture by decompression.
出处
《物探化探计算技术》
CAS
CSCD
2001年第2期150-154,共5页
Computing Techniques For Geophysical and Geochemical Exploration
关键词
离散余弦变换
矢量量化
图像压缩
图像解压
DCT变换
压缩算法
discrete cosine transform
vector quantization
self organizing feature map
image compression
image decompression
