Geometric Estimation of FoV for Vision Modeling in DHM 2009-01-2294
This paper presents a geometric method for estimating the filed-of-view (FoV) for monocular and binocular vision using a 3D scanned human face. The traditional vision-cone approach offers simplicity but may lead to spurious inferences during ergonomic evaluations involving peripheral vision. FoV is asymmetric due to unrestricted zygomatic vision and restrictions on nasal side. Thus its accurate modeling is important for effective simulation of visual performance. Given a tessellated face model and the location of the pupil, the present work computes the FoV edges on the face with respect to the pupil. The algorithm is divided in two steps, visible triangle computation and silhouette edge detection for FoV estimation. A qualitative comparison of samples with the standard perimetry data and right circular cones used in DHM shows that the FoV is very similar to the physiological visual-field maps and the cones used in DHM are conservative. The proposed approach has been extended to gaze dependant field-of-view computation which is difficult to obtain experimentally due to involuntary eye movements in subjects. The gaze-dependant FoV is found to be significantly different from normal monocular and binocular FoV. Thus the present computational FoV paradigm has the potential of active vision integration in performance evaluations using DHM's.