...期望最大化算法；场景分类；K-L 散度 文献标识码：A [gap=1010]ey Words: Multi-Instance learning; Information Bottleneck; Gaussian Mixed Model; Expectation Maximization; scene classification; Kullback-Leibler Divergence ..
基于36个网页-相关网页
该算法借用了雷达信号处理中CLEAN算法的思想,其性能明显优于期望最大化算法(Expectation Maximization Algorithm, EM)。进一步针对跟踪滤波中运动目标模型的不确定性,提出了多模型PHD滤波算法。
基于12个网页-相关网页
隐马尔可夫模型参数通过期望最大化算法(EM)来估计。
这里，期望最大化算法既用来处理丢失值又用来估计模型参数。
We resort to expectation maximization (EM) algorithm for both the estimation of model parameters and the coping with missing values.
在此基础上给出期望最大化算法图像恢复结果，并对恢复结果做出分析。
The deconvolved results with the compensated data and the original image data by expectation maximization (EM) algorithm for reducing the effect of out of focus light were compared respectively.
应用推荐