3D ear modeling and recognition from video sequences using shape from shading