A Back-Propagation Neural Network Based on a Hybrid Genetic Algorithm and Particle Swarm Optimization for Image Compression

摘要：

In this paper, an improved approach integrating genetic algorithm and adaptive particle swarm optimization with feed forward neural networks for image compression is proposed. The hybrid genetic algorithm with a novel mutation strategy and particle swarm optimization is used to train the neural network to near global optimum weights and thresholds at first. Then the network is trained with gradient descending learning algorithm to obtain the optimal network parameters. Then, the trained network is applied to the image compression. Results show that at the same compression rate the application of optimized neural network in image compression will achieve better image quality compared with the application of traditional neural network.

关键词： image compression BP neural network PSO evolutionary strategy

作者: Han Feng Man Tang Jie Qi

作者单位: College of Information Science and Technology Donghua University Shanghai, China

会议类型: 国际会议

会议名称: 2011 4th International Congress on Image and Signal Processing(第四届图像与信号处理国际学术会议 CISP 2011)

会议地点: 上海

会议语种:英文

页码: 1336-1339

在线出版日期: 2011-10-15（万方平台首次上网日期，不代表论文的发表时间）