Abstract:
Aiming at the problems of much iteration number and long training time for BP’s application in image compression, a Quantum BP(QBP) network algorithm is proposed in this paper. By using a neuronal model with quantum input and output, combined with the theory of BP in image compression and the plural BP algorithm, a model for image impression with 3-layer QBP network is built, which implements image compression and image reconstruction. Simulation results show that QBP network can obtain the reconstructed images with better quantity compared with BP network.
Key words:
Quantum Neural Network(QNN),
iteration number,
Quantum BP(QBP) network,
plural BP algorithm,
image compression,
convergence rate
摘要: 针对BP网络在图像压缩应用中迭代次数多及训练时间长的问题,设计具有量子输入和输出的神经元模型,结合BP网络在图像压缩中的原理,利用复数BP算法,构建一种用于图像压缩的3层量子BP网络(QBP),实现图像压缩与图像重建。仿真结果表明,与BP网络相比,QBP网络能获得更好的重建图像质量,且迭代次数较少。
关键词:
量子神经网络,
迭代次数,
量子BP网络,
复数BP算法,
图像压缩,
收敛速度
CLC Number:
ZUO Xian-Gang, ZHANG Zhi-Xia. Image Compression Method Based on Quantum BP Network[J]. Computer Engineering, 2012, 38(13): 205-207,211.
左现刚, 张志霞. 一种基于量子BP网络的图像压缩方法[J]. 计算机工程, 2012, 38(13): 205-207,211.