Occupant labeling for airbag-related applications

摘要

A segmented image (42) of an occupant in a vehicle is input to a k-metric module (60) which generates upper and lower torso metrics with occupant classification. An iterative parameter estimator (66) generates upper and lower torso parameters from the metrics. A pixel classifier (72) generates upper torso image from the generated torso parameters in order to obtain occupant characteristics. Independent claims are also included for the following: (1) occupant tracking and airbag deployment system; and (2) upper torso image identifying method.