Three Dimensional Object Recognition Using a Complex Autoregressive Model

摘要

Based on an autoregressive model, Complex Partial Correlation (CPARCOR) featuresare known to provide exceptional Position, Scale, and Rotation Invariant (PSRI) properties for planar 2-Dimensional (2-D) object recognition. Although autogressive models have been successfully applied to numerous spatio-temporal recognition tasks, the effects of out-of-plane image rotations were never considered. This study investigates application of the CPAR-COR model to a five class problem of nonplanar 2-D views of 3-D objects. Recognition based on CPAR-COR features is evaluated using a Template Matching algorithm, two K-Nearest-Neighbor (KNN) classifiers, and a Hidden Markov Model (HMM). Direct comparisons to recognition based on Fourier features are made. Results indicate that the CPAR-COR model parameters provide useful shape-features for recognition of out-of-plane rotations. Displaying exceptional PSRI properties, the features are shown capable of classification by simple nonadaptive recognition schemes. Relatively successful results are obtained for a variety of tests. The advantage of classification by a multiple-look technique over the traditional single-look method is clearly demonstrated. Feature space crowding is noted as the cause of unusual recognition rates for occluded-view tests. Although general trends are noted, optimal model order and selection of CPARCOR versus Fourier features are considered application dependent. Autoregressive model, Multiple-look classification, Occluded-view recognition, KNN Classifiers.