New Multifractal Techniques for Pattern Analysis in Biomedical Images

This book develops and evaluates different multi-fractal techniques for detection, classification and analysis of medical images. Multi-fractal analysis has a better way of handling irregularities compared to the classical approach as it extracts information directly from the singularity, and the features obtained are used for the descriptions and analysis of the image patterns.The uniqueness about this book is that it uses multi-fractal as a tool for solving classification problems in biomedical images since the scaling properties of most natural images exhibit self-similarity properties that are very common in most medical diseases. This research work studies the structures of emphysema CT images and extracts the statistical self-similarity features characterized by the Holder exponent to develop new descriptors for pattern classification. The results achieved in the book could be very useful in academic environments, research institutes, and also to assist professionals in the field of medicine.