Signal and image processing with neural networks : a C++ sourcebook
Timothy Masters (Author)
This unique book/disk set is the only guide you need to learn practical, proven techniques for integrating traditional signal/image-processing methods with modern neural networks. It provides thorough, clear, and highly readable coverage of complex-domain neural networks, which are generally superior to the usual real domain models when processing complex data. Signal and Image Processing with Neural Networks presents the only detailed descriptions available in print of standard multiple-layer feedforward networks generalized to the complex domain. Always stressing the practical and the intuitive, this book/disk set will enable you to perform all relevant techniques and procedures. C++ subroutines are provided for all techniques, and the discussion of data preparation for neural networks focuses on special requirements for signal and image processing. Techniques presented include gabor transforms in one and two dimensions; wavelets, focusing on the Morlet wavelet, in one and two dimensions; object identification based on shape via Fourier analysis of the perimeter and shape moments; image classification via tone and texture variables; fast transforms in the frequency domain; integration of these techniques with modern neural networks; and much more
Print Book, English, ©1994
J. Wiley, New York, ©1994