Estimating head's measurements from front-view head and shoulders images

The automated process of estimating a human head's measurements from a front-view head and shoulders image/video sequences is crucial for applications such as automated face identification, model-based video coding, and database indexing and browsing. This paper introduces a new model-based algorithm for estimating the measurements of a human head, including the head center, tilt angle, length and width. The algorithm is based on modeling the human head in a front-view head and shoulders image as an ellipse. The algorithm estimates the parameters of the ellipse which best fits the head view in the image and uses these parameters to estimate the head center, the tilt angle, the length, and width. The algorithm has been implemented and tested with over twelve hundred images, and simulation results indicate that the algorithm is robust to variations in subject head shape, age, and motion such as tilting and nodding of the head.