%0 Journal Article
%T Neural Network Competitive Learning Algorithms for Vector Quantization<br>用于矢量量化的神经网络竞争学习算法
%A Xu Yong
%A Chen Hexin
%A Dai Yisong
%A <br>徐勇
%A 陈贺新
%A 戴逸松
%J 中国图象图形学报
%D 1997
%I 
%X A fast VQ image coding method based on humans visual attribution and applying wavelet tree structure is proposed in this paper, naming tree structure fast VQ coding. After the characteristics of VQ was analyzed, a statistic method generating a codebook was designed, and a fast VQ coding method was represented. This method can efficiently remove correlation in image data, obtaining a low transmission bit stream. The apparent advantage of the method is to estab lish statistics codebooks for various image data, and each treating need not generate codebook s with a high coding efficient achieved, while it can apply to any opportunity with wavelet transformation used in various image and signal data compression. The experimental results show: The fast VQ coding method proposed in this paper can achieve a compression ratio of 40 with a PSNR of 36.21dB,while its total performance is superior to other methods, and a real time compression using this method may be implemented.
%K Neural network
%K Competitive learning
%K Vector quantization
%K Neuron underutilization
%K Algorithm<br>神经网络
%K 竞争学习
%K 矢量量化
%K 算法
%U http://www.alljournals.cn/get_abstract_url.aspx?pcid=5B3AB970F71A803DEACDC0559115BFCF0A068CD97DD29835&cid=8240383F08CE46C8B05036380D75B607&jid=D06194629680C940ACE75262F54B9D85&aid=E0E89D82A8CE04209F6F0B6F8D5719E3&yid=5370399DC954B911&vid=0B39A22176CE99FB&iid=59906B3B2830C2C5&sid=E23857687CD2EE51&eid=46CB56AABC2765FF&journal_id=1006-8961&journal_name=中国图象图形学报&referenced_num=0&reference_num=0