Multispectral Imaging: Application to Pathology: Machine vision and texture analysis for the automatic identification of tissue patterns in prostatic tumours

This book contains the work conducted by Dr. M. Ali Roula during his PhD research project. It is concerned with an investigation into improving the classification of biopsies for prostate cancer diagnosis. The concept of texture analysis is explored from different angles. By making use of the recent development of technologies such as high-throughput Liquid Crystal Tuneable Filters enabling capture of narrow band multispectral images of tissue sample. When analysing the new multispectral images, it was found that the visible spectrum contains significant information, which was not captured using standard methods such as grey level and RGB. In this book, nuclei chromatin texture is also explore using the novel multispectral approach. Results showed a significant improvement in identifying different cell groups when compared with standard RGB methods. Finally, the analysis of nuclei is also extended using 3D texture measurements with potential application in confocal laser scanned microscopy images.