275 / 1971-01-01 00:00:00

An Improved Camshift Algorithm For Face Tracking

Camshift,AdaBoost,face tracking,Kalman filtering

Camshift is an effective algorithm for real time dynamic target tracking applications, which only finds target manually and is sensitive to occlusion.To resolve the problem, an improved Camshift algorithm is proposed to seek the initial search window automatically by combined with Adaboost face detection. Besides,skin color detection is added to the back-projection image to increase the weight of face region.Once the target is judged lost, Kalman filtering is utilized to find it again and the improved Camshift keeps on tracking the target continuously.The results prove that the proposed method is more robust than the traditional Camshift and give better tracking performance than some other improved methods.