采用线性回归树作为弱分类器,再由多个弱线性回归树组成强分类器以提高分类能力.从而在不增加特征数的情况下,通过线性回归树将特征进行自动组合以提高弱分类器的分类能力:相应地可以用更少的弱分类器组成分类能力更强的强分类器完成分类任务.对PETS2006视频序列以及公交车内视频序列的目标进行了检测,证明其具有较强的分类能力和较好的检测效果.
A new boosting based learning method was proposed, in which the linear regress tree was sed to automatically combine multiple features in weak leaning process. The final strong classifier was composed of several weak linear regress trees. This leads to a better strong classifier, which conists of fewer weak classifiers and features. And the result ssytem can keep computation cost low. Exerimental results on PETS2006 dataset show the strong classification capability and high detection rate.