An Improved AdaBoost face detection algorithm based on the weighting parameters of weak classifier

摘要

Weighting parameters are introduced to ensure the weak classifier that comes with the False Rejection Rate (FRR) to significantly reduce the False Acceptance Rate (FAR). Knowing that the Haar-Like features redundancy, the most effective combination of features is chosen from all the features upon the completion of the classifier training, aiming to improve the speed and rate of face recognition. The results show that the improved AdaBoost algorithm saw an improved recognition rate of 15% compared to the traditional algorithm, where the video image sequence presented an average face recognition rate of 21.5ms/frame, being able to meet the requirements of real-time face detection.