Geometric neurocomputing for pattern recognition and pose estimation

Presents a geometric neurocomputing approach for 3D pose recognition using the framework of geometric Clifford algebras. The type of geometric problems like pattern recognition and 3D pose recognition can be very efficiently handled using geometric neural networks. Our experimental part shows the application of generalized Clifford moments for pattern recognition and 3D pose estimation of rigid objects using visual information captured by a trinocular head.