Human Recognition Based on Random Forest Classifier of HOG

Aimed at the question of low recognition ratio and high error rate,a random forest classifier is put forward based on HOG which combines the good representation of the appearance and shape of partial image of HOG algorithm and robust target classification performance and effect of the random forest classifier.Besides,it has better robustness by comparing with binary tree,AdaBoost and SVM classifier.At last,the random forest classifier of HOG gains the effective verification in a complex scene.