D.B.L. Bong, N. Rajaee, S.S. Ngu, and L.C. Kho (Malaysia)
Face detection, tracking, HSI, Kalman Filter
Face detection and tracking is important in many applications such as in security video surveillance, face recognition, face image database management and human computer interface. This paper presents a novel approach in detecting and tracking faces using a hybrid system of HSI (hue, saturation, intensity) color skin model and Kalman Filter. Range of values for hue, saturation and intensity of skin color is used to locate and identify potential face candidates in video sequences. Areas that fall within the desired HSI color space undergo object segmentation process utilizing Euler Number to label and segregate the face candidates. Presence of human face for each potential candidate is determined by a threshold function using Euler Number. Next, Kalman Filter is used to continuously track the movement of the faces in subsequent video images. Experiment showed good results in detecting and tracking frontal faces in color video sequences.
Important Links:
Go Back
