快速分形图像压缩方法及其VLSI实现

Method of high-speed fractal image compression processing and its VLSI implementation

针对分形编码技术解码迅速但编码时间较长的特点,提出了一种结合kohonen神经网络(KNN网)和多尺度分析的分形编码算法,并在此基础上给出了基于该方法的硬件实现方案．该方法可大大加速图像编码过程而使压缩率和图像质量仅有微小变化．实验结果显示,其压缩速度比全搜索算法提高了131倍．

In light of the character of the fractal coding technology which allows fast decoding but spends long encoding times, a new method connected kohonen neural nets with muhiresolution analysis is presented, and an efficient VLSI architecture is proposed based on it. This method can greatly speed up image coding process with little worse image quality and compression ratio. The experiment results show that its compression speed is 131 times as quick as the exhaustive search method.