Fast convergence of regularised Region-based Mixture of Gaussians for dynamic background modelling

dblp(2015)

引用 16|浏览34
暂无评分
摘要
•Derivation of O(1/k2) rate of convergence for online gradient method with momentum.•Novel foreground detection algorithm for dynamic background modelling.•Fast convergence of mixtures demonstrated on various sets of simulated data.•Superior performance compared to other state-of-the-art algorithms on real videos.
更多
查看译文
关键词
Momentum,Convergence rate,Generative models,Region-based Mixture of Gaussians,Dynamic background modelling
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要