Improved Face Detection Algorithm Based on Adaboost

In view of the training time-consuming shortcoming of the conventional AdaBoost algorithm in face detection, in this paper, a new algorithm was presented combining effectively the optimizing rect-features and weak classifier learning algorithm, which can largely improve the hit-rate and decrease the train time. Optimized rect-feature means that when searching rect-feature we can establish a growth step length of the rect-feature and reduce its features.And the new weak classifier training method is seeking the weak classifier error rate directly, which can avoid the iterative training, the statistic probability distribution and any other time-consuming process. The experimental result showed that the algorithm mentioned in this paper, reduces training time cost greatly compared with conventional AdaBoost algorithm. In addition, it speeds up weak classifier and improves the detection speed on the premise of high detection accuracy.