Customer Reviews

Unsupervised Adaptive Filtering, Volume 1: Blind Source Separation

5 star:		(2)
4 star:		(0)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

This book is essential for anyone who wishes to stay informed about new developments in the very important field of signal processing. In the introduction, Simon Haykin (of Canada's McMaster University) explains the classification of supervised and unsupervised forms of adaptive filtering and the three fundamental approaches to unsupervised adaptive filtering. He also explains that blind deconvolution (the subject of volume 2) involves inverse problems with both similarities and differences vis-a-vis blind source separation. In an important way, volume one is an indispensable introduction to the second volume.

The contributors to volume one are an impressive collection from both academic research and the private sector. Haykin has assembled an international team from institutions and private laboratories in France, Japan and the United States. Their contributions are both theoretical and practical. The chapters include references to important literature in signal processing, communications and control, to facilitate further research.

The first seven chapters are information-theoretic approaches. The final two chapters present important practical issues in solving problems with delayed or convolved mixtures of independent source signals. Chapter Nine (by R. Lambert and C. Nikias) is especially useful, as it presents a powerful and novel tool for separating real-life speech.

This book is a significant contribution to the signal processing literature and deserves a place on the shelf of anyone interested in adaptive filtering.

Succinct and clear explanations, broad coverages and the most recent examples like functional MRI.