A Multimodal Fruition Model for Graphical Contents in Ancient Books