The background samples are chosen by thresholding inter-frame differences, and the Gaussian kerneldensity estimation is used to estimate the probability density functionof background intensity.
通过相隔固定的帧差值阅值化得到背景样本值,并采用高斯核密度估计方法计算背景灰度的概率密度函数。
2
In this paper, a new kernel estimator of multivariate density is proposed by using a univariate kernelfunction.