Progress in Neural Networks, Volume 4: Machine Vision

Progress in Neural Networks, Volume 4: Machine Vision

O.M. Omidvar and R. Mohan (Eds)Ablex Publishing Corporation1997305 pp.ISBN 0-89391-967-5£39.95 Hard cover

Keywords: Neural networks, Machine vision

This volume contains eight papers that offer a new framework for characterising and solving tasks that cannot be achieved using conventional computing techniques. The papers illustrate a number of ways of formulating and analysing problems in the computational framework of neural networks. The rich cross-fertilisation between neuro-psychology and computer vision is also highlighted within the different chapters.

The eight chapters cover a variety of topics ranging from low-level to high-level vision and ways of applying them to different neural network architectures. Chapters 1 and 2 address "receptive field calculus" and "visual reconstruction and data fusion", respectively. The "visual perception of translational and rotational motion" and a "neural network for recovering observer motion" are discussed in chapters 3 and 4.

The following three chapters address different topics in image segmentation. These are "image segmentation using a multilayer Kohonen's self organising feature map", "illusory contours and image segmentation: neural network architectures" and "grouping and segmentation using neural networks:applications to edge detection, texture segmentation and face recognition". The final chapter addresses "object recognition using constraint satisfaction networks".

Overall, this book is suitable for engineers, computer scientists,neuro-psychologists and anyone who is interested, or involved in, machine vision. This volume is both informative and interesting.