Journal of Computer Applications ›› 2011, Vol. 31 ›› Issue (03): 706-709.DOI: 10.3724/SP.J.1087.2011.00706

• Graphics and image technology • Previous Articles     Next Articles

Study on adaptive ability of Gaussian mixture background model

ZHANG Yun-chu1,SONG Shi-jun2,ZHANG Ru-min3,HAO Jian-lin3   

  1. 1. School of Information and Electrical Engineering, Shandong Jianzhu University, Jinan Shandong 250101, China; Shandong Provincial Key Laboratory of Intelligent Buildings Technology, Shandong Jianzhu University, Jinan Shandong 250101, China
    2. School of Mechanical and Electronic Engineering, Shandong Jianzhu University, Jinan Shandong 250101, China
    3. School of Information and Electrical Engineering, Shandong Jianzhu University, Jinan Shandong 250101, China
  • Received:2010-09-14 Revised:2010-11-14 Online:2011-03-03 Published:2011-03-01
  • Contact: ZHANG Yun-chu

高斯混合背景模型的适应能力研究

张运楚1,宋世军2,张汝敏3,郝建林3   

  1. 1. 山东建筑大学 信息与电气工程学院,济南250101; 山东建筑大学 山东省智能建筑技术重点实验室,济南250101
    2. 山东建筑大学 机电工程学院,济南250101
    3. 山东建筑大学 信息与电气工程学院,济南250101
  • 通讯作者: 张运楚
  • 作者简介:张运楚(1968-),男,山东曹县人,教授,博士,主要研究方向:图像处理、计算机视觉、模式识别、智能建筑;宋世军(1965-),男,山东德州人,教授,博士,主要研究方向:图像处理、计算机视觉、机械工程;张汝敏(1983-),女,山东聊城人,硕士研究生,主要研究方向:图像处理、计算机视觉;郝建林(1983-),男,山东烟台人,硕士研究生,主要研究方向:图像处理、计算机视觉。

Abstract: Gaussian mixture background model is an online parameterized statistical model, and the presenting way of pixel sample pattern observed from time window has great influence on the model's learning result. According to the characteristics of dynamic background changes, issues such as the stability and plasticity of modal learning, the modal residual and activation, which affected the model's adaptive ability, were studied. The simulation results show that Gaussian mixture background model has a robust selective adaptability to gradual change, but a limited adaptability to transient variation of background configuration provided by modal residual and activation mechanism.

Key words: Gaussian mixture background model, motion segmentation, adaptive ability, modal residual

摘要: 高斯混合背景模型是一种参数化统计模型,观察时间窗内像素样本模式呈现规律决定了背景模型的学习结果。针对背景动态变化的特点,研究了影响背景模型适应能力的模态稳定性与可塑性、模态残留与激活问题。仿真实验表明高斯混合背景模型具有较强的渐变选择性适应能力,而模态残留与激活机制为模型提供了有限的背景结构短时变化适应能力。

关键词: 高斯混合背景模型, 运动分割, 适应能力, 模态残留

CLC Number: