基于Dirichlet过程的无线视频码率变化识别算法

摘要

研究了通过对终端视频帧质量的聚类分析来识别无线视频传输中码率变化的方法,以便为无线视频传输过程中视频码率自适应调整提供参考依据.针对经典模糊C均值(FCM)算法和K均值(K-means)算法需要设定聚类数目的问题,提出一种基于荻利克雷过程(DP)的FCM算法——DP-FCM算法.该算法将Dirichlet过程和FCM算法相结合,由视频帧信息权重峰值信噪比(IWPSNR)值使用DP过程混合模型模拟估计出聚类数目,然后进行FCM模糊聚类,通过设定合理的阈值,合并聚类结果相似项,完成视频帧的聚类,从而实现视频传输码率变化的识别.以LIVE视频库为试验数据源,对该算法进行了性能测试.试验结果表明,DP-FCM算法能够在无需设定聚类数目的前提下实现视频传输码率变化的分类识别.%The recognition of the changes of code rate of wireless video transmission was studied through clustering analysis of the quality of terminal video frame to provide technical references for adaptive adjusting the code rate during wireless video transmission.In view of the problem that classical recognition algorithms of fuzzy c-means (FCM) and K-means need setting cluster number in advance,the study proposed a Dirichlet process (DP) based FCM algorithm,called DP-FCM algorithm.The proposed DP-FCM algorithm combines the Dirichlet process with the FCM algorithm,and estimates the number of clusters using the DP mixture model and video frame quality.Then fuzzy c-means clustering for video frame quality is performed.The similar clustering results are merged by setting a reasonable threshold.Finally,video transmission rates are recognized by clustering video frames.The results of the experiment conducted on the LIVE database demonstrate that the classification and identification of video transmission rate can be achieved by the DP-FCM algorithm without setting the number of clusters.