Abstract. In this paper we present a unique approach to personal identification which utilized finger surface features as a biometric identifier. Finger surface features are extracted from dense range images of an individual's hand. The shape index (a curvature-based surface representation) is used to represent the surfaces of the index, middle, and ring fingers of an individual. This representation is used along with a correlation coefficient based matcher to determine similarity. Our experiments make use of data from 223 subjects possessing a 16 week time lapse between collections. We examine the performance of individual finger surfaces in a verification context as well as the performance when using the three finger surfaces in conjunction. We present the results of our experiments, which indicate that this approach performs well for a first-of-its-kind biometric technique.