摘要
本文介绍了神经网络在图象压缩领域的应用,并且对前馈神经网络模型及其学习算法进行改进。在改进的模型中,隐层的每个神经元都附加上一个基于Logistic映射的伴随神经元,能够产生混沌现象。同时论述了这种混沌神经网络用于图像压缩的模型、原理、算法及关键技术。并通过仿真实验将这种模型应用于图像压缩中,其中的关键是算法,激活函数和压缩率等参数的选择。在固定压缩率和激活函数的条件下,将这种模型和最常用于图像压缩的标准BP网络模型进行了比较,结果表明这种混沌神经网络在压缩率,失真率等方面性能更好,从而使得重建图像效果更好。
This paper has presented the application of neural network in image compression, and the model and algorithm of a feed-forward network is improved. In the new model, every nerve cell of the hidden-layer is adding a concomitant cell which is brought by Logistic mapping, and it has a chaos characteristic. The model, theory, algorithm and key technology of the Chaos neural network was discussed when it was applied in image compression in this paper. The network can use in image compression by simulation experiment. When the compression ratio and activation function is fixed, the chaos neural network is improvement comparing to common BP network in image compression. So the effect of rebuild image is better.
出处
《仪器仪表用户》
2008年第4期61-62,共2页
Instrumentation
