Pursuing mechanical part feature recognition through the isolation of 3D features in organic shapes

The successful extraction of 3D features in mechanical parts has always been a challenging task and has yielded mixed results. Extracting features from organic shapes however is even more difficult. This is due to the fact that they are defined by both gradual and abrupt changes in surface curvature. The term curvature is explained in detail in section 4. Learning how to recognize organic shapes may give insights into better ways of performing feature recognition on mechanical parts. Determining the exact values of curvature, based on the underlying parameters can prove to be quite difficult. Curvature can be a good tool to identify features as most of the features are areas of slowly changing curvature bounded by sudden changes in curvature. The benefits of developing a generic algorithm that picks out curvature, and hence the organic features, are quite huge. This paper explains one approach taken to accomplish this task. This paper studies characteristics of the watershed algorithm[1] when applied to the features on bones. This algorithm is used to isolate features based on curvature gradients. This paper uses the knowledge from the field of anthropology and medicine to explore the sensitivity factor analysis of the watershed and its effectiveness in extracting the features on bones. The paper also compares the differences between the anatomists definition of a feature and the algorithm interpretation of the same feature.