Improved Viola-Jones Face Recognition Using Tracking
Optimization of accuracy has been an important study in the study of face recognition for 2 decades. Varied facial features and changing poses in a short amount of time make optimizing accuracy even more complicated. This study aims to improve the accuracy of face recognition of the Viola-Jones method on moving objects by adding a tracking algorithm. The tracking algorithm used in this study is the Continuously Adaptive Mean Shift (Camshift) algorithm. This algorithm is the development of the Mean Shift algorithm which continuously adapts or adjusts to the color probability distribution that is always changing with each change of frame of a video sequence. The addition of tracking significantly increases accuracy compared without tracking by 96%.