FIG. 6B illustrates an embodiment of an example image 650 captured from the computing devices of FIG. 6A. The image 650 can be either a 2D or 3D image, such as a raw JPEG (2D) image or raw PLY (3D) image having point cloud data. Image 650 depicts a driver and several types of identifiable driver behaviors, e.g., items of interest, that can be determined from points or pixels of the image. For example, point 654 of FIG. 6B, depicting the driver's forehead, can correspond to 3D point 506 of FIGS. 5A and 5B. Similarly, point 656 of FIG. 6B, depicting the driver's cheek or lip area, can correspond to 3D point 504 of FIGS. 5A and 5B. As described herein, the points 654 and 656 may be items of interest that may be used for identification (e.g., facial recognition to determine the position of the driver) or used by classification of driver behavior, or determination or development of a related risk factor value. For example, image 650 includes other items of interest, for example, as identified by points 660 and 662. Point 660 relates to the driver's hand, which, as shown, is on the steering wheel of the vehicle. In certain embodiments, the identifications of a driver's hand on the steering wheel could indicate safe driving, and thus, a risk value associated with the driver may be improved (e.g., a lowering the risk value). Point 662, however, relates use of a mobile phone. Accordingly, in certain embodiments, the identifications of use of a mobile phone could indicate dangerous or risky driving, and thus, the risk value associated with the driver may be adjusted accordingly (e.g., increasing the risk value).
