We describe a ne wapproach for creating ocncise high-level gererative models form range images or other approximate representations of real objects.using dat from a variety of acquistin techique s and a user-deffind class of modles,our mehtod produces a compact object representation that is intuitive and easy to edit.The algorithm has two inter-related phases: recognition,which chooses an appropriate model within a user-specified hierarchy,and parameter estimation,which adjusts the model to best fit the data.Since the approach is modle-based,it is relatively insensitive to noise and missing data.We describe practical heuristics for automatically making tradeoffs between simplicity and accuracy to select the best model in a given hierarchy.We also describe a general and efficient technique for optimizing a model by refining its ocnstituent curves.We demonstrate our approach for model recovery suign both real and synthetic data and several generative mdoel hierarchies.
展开▼