Invariant pattern recognition with higher-order neural networks.
Zhengquan. He · DR-NTU (Nanyang Technological University) · 1999
Translation, rotation (in plane) and scale invariant pattern recognition is a high-order recognition problem encountered frequently in real-world applications. But neither traditional image process/pattern recognition algorithms nor artificial neural networks have yet provided satisfactory solutions for this problem after years of study. Recent research has shown that a higher order neural networks (HONNs) of order three with built-in invariances can effectively achieve invariant pattern recognition.