基于神经网络的分形图像压缩与解压缩  

Fractal Image Compression and Decompression Based on Neural Network

作　　者：李明亮[1] 王祖朝[1] 母景琴[1] 李聚光[2] 

机构地区：[1]中国地质大学信息工程学院,北京100083 [2]石家庄经济学院,河北石家庄050031

出　　处：《河南科技大学学报（自然科学版）》2006年第5期32-34,37,共4页Journal of Henan University of Science And Technology:Natural Science

基　　金：河北省科学技术研究与发展基金项目(05213510D);河北省科技厅基金项目(02213508D)

摘　　要：图像压缩技术中,分形图像压缩与解压缩具有较高的压缩率和低损耗率,但基于迭代函数系统的自动分形图像编码方法计算量大,因此限制了其实际应用。人工神经网络具有学习、记忆、识别及推理等特点,为分形图像压缩解压缩提供了一个全新的解决方案。提出使用神经网络方法以并行方式完成分形图像压缩与解压缩的大量计算,模拟结果显示,与基本的自动分形图像编码方法相比,在基本保证重建图像质量的前提下,运算时间和码率均下降,显示出神经网络技术应用于分形图像压缩与解压缩中的可行性及高效性。In image compression technologies,fractal image compression/decompression have the advantages of a high compression ratio and a low loss ratio.However,it requires a great deal of computation,which limits its applications.The artificial neural network(ANN) is a new project that carries on fractal image compression/decompression with characteristics of learning,memory and reasoning.This paper uses neural networks to perform a large number of computations in fractal image compression and decompression in parallel.The simulation results show that the quality of images generated by neural networks is similar to that produced by basic automation fractal image coding,however,both operation time and code rate are decline.So the neural network technology is useful and efficient when applied to fractal image compression and decompression.

关 键 词：分形图像 压缩/解压缩 神经网络 并行处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]