| 基于神经网络的VBR视频流量建模及预测 |
| |
| 引用本文： | 王以宁,;李兴华,;迟学芬.基于神经网络的VBR视频流量建模及预测[J].长春邮电学院学报,2008(5):509-513. |
| |
| 作者姓名： | 王以宁 ;李兴华 ;迟学芬 |
| |
| 作者单位： | [1]东北师范大学传媒科学学院,长春130117; [2]中国电信股份有限公司珠海分公司,广东珠海519000; [3]吉林大学通信工程学院,长春130012 |
| |
| 基金项目： | 吉林省自然科学基金资助项目（20060525） |
| |
| 摘 要： | 为解决视频流量预测问题，结合神经网络和小波技术建模IP（Internet Protoc01）网络视频流，提出了利用神经网络预测尺度因子的预测算法。对可变比特率的压缩视频流完成小波分解，得出尺度因子。通过对尺度因子的预测和小波重建，完成视频流量预测。尺度因子的归-化特性简化丁神经网络处理过程。对真实VBR（Variable BitRate）视频流的流量预测实验表明，模型对IP网络普遍应用的高压缩比视频流具有良好的预测能力。
|
| 关 键 词： | 神经网络 视频流量 预测 尺度因子 |
VBR Video Traffic Modeling and Prediction Based on Neural Network |
| |
| Affiliation: | WANG Yi-ning , LI Xing-hua, CHI Xue-fen (1. School of Media Science, Northeast Normal University, Changchun 130117, China ; 2. Zhuhai Branch, China Telecom Corporation Limited, Zhuhai 519000, China; 3. College of Communication Engeering, Jilin University, Changchun 130012, China) |
| |
| Abstract: | Combining neural network and wavelet technologies, establishes an IP （Internet Protocol ） network video streaming model, in which a prediction algorithm based on neural network is proposed to predict the scale factor. The scale factor is obtained by completing the wavelet decomposition to VBR （Variable Bit Rate） video traffic. Video traffic prediction is implemented by forecasting the scale factor and completing wavelet reconstruction. The scale factor is always less than 1, this characteristic makes computing process of neural network simple. The experiments based on real VBR video traffic show that this method has good performance on forecasting IP network video flow, which usually has high compress ratio. |
| |
| Keywords: | neural network video traffic prediction scale factor |
| 本文献已被 维普 等数据库收录！ |
|
