Springer Human Face Recognition Third-Order Neural Networks

Explore the advanced application of High-order synthetic neural network technology in this technical volume from Springer. This book examines how these networks facilitate transformation-invariant recognition of complex visual patterns. High-order networks offer a distinct advantage by requiring very little training data, which leads to significantly shorter training times. While these methods have already proven successful in performing transformation-invariant recognition of simple visual patterns with high recognition rates, this work moves toward addressing more practical problems. Specifically, the research transitions from binary patterns to the challenges of grayscale recognition. This text serves as a specialized resource for those studying the evolution of synthetic neural networks and their ability to handle complex, real-world visual data through advanced mathematical modeling.