Optical Flow Estimation Using Pulse Coupled Neural Networks

Most of the ways for Optical Flow estimation in computer vision depend on static algorithms or dynamic algorithms which need some kind of training. The subject of this work is to use a model of biological neural networks called Pulse Coupled Neural Networks, which try to model biological neuronal networks in a more accurate manner, to implement a system for Optical Flow estimation which needs no training and is suitable for the use with both artificial and natural images. Beginning with a introduction to the subject and the problems of Optical Flow estimation as well as to Pulse Coupled Neural Networks this book suggests a multi layer architecture which should enable the reader to implement a system on his own.