Bayesian estimation of Dirichlet mixture model with variational inference
作者：
Highlights：
• An analytically tractable solution for Bayesian estimation of the Dirichlet mixture model.
• Relative convexity of the multivariate log-inverse-gamma function is proved and utilized.
• The free energy function is approximated by a single lower-bound to guarantee convergence.
• The method outperforms the ML based method and the VI based method, moreover, it is comparable to the sampling based method.
• The performances are demonstrated with important multimedia signal processing applications.
摘要
Highlights•An analytically tractable solution for Bayesian estimation of the Dirichlet mixture model.•Relative convexity of the multivariate log-inverse-gamma function is proved and utilized.•The free energy function is approximated by a single lower-bound to guarantee convergence.•The method outperforms the ML based method and the VI based method, moreover, it is comparable to the sampling based method.•The performances are demonstrated with important multimedia signal processing applications.
论文关键词：Bayesian estimation,Variational inference,Extended factorized approximation,Relative convexity,Dirichlet distribution,Gamma prior,Mixture modeling,LSF quantization,Multiview depth image enhancement
论文评审过程：Received 13 June 2013, Revised 1 February 2014, Accepted 1 April 2014, Available online 12 April 2014.
论文官网地址：https://doi.org/10.1016/j.patcog.2014.04.002
