This paper explores face recognition from a representation based on features extracted from range images. Depth and curvature features have several advantages over more traditional intensity based features. Specifically, curvature descriptors 1) have the potential for higher accuracy in describing surface based events, 2) are better suited to describe properties of the face in areas such as the cheeks, forehead, and chin, and 3) are viewpoint invariant. Faces are represented in terms of a vector of feature descriptors. Comparison between two faces is made based on thier relationship in the feature space. We provide detailed analysis of the accuracy and discrimination of the particular features extracted, and of the effectiveness of the recognition system for our test database of 24 faces. Results are very promising. In many cases it is shown that feature accuracy is limited more by surface resolution than by the extraction process. Recognition rates in our experiments are in the range of 80% to 100%.