| Recent hardware advances in personal computing power have spurred the use of 3D models. However, modeling realistic human faces is still a challenging task. There is a need for an input device, analog to a camera for 2D data to enable capturing of 3D surface information for human faces.; This study focuses on the design of a 3D digitizer for human faces. The design accounts for the natural human movements. Two methods of 3D extraction from 2D images, phase-shift and direct-structured light, are investigated in which color is used to increase the amount of information that can be obtained from a single image.; The phase system method was found to be inadequate for this task. The presence of color crosstalk in both the projection system and the detector (digital cameras) introduce ripples that require heavy filtering of the result, negating the main advantage of producing high resolution output.; A direct structured light method using RGB color plus one other distinct color (white in this case) for exact identification was designed. Experiments on real human faces showed that the system could produce a 3D surface profile of human faces with sub-millimeter accuracy. Accuracy analysis found the presence of some distortion effects. A simple calibration procedure was designed to minimize the distortion effect. Simulations were performed to verify that the desired accuracy could be achieved by the direct structured light method. |
