Medical Image Analysis and Informatics: Computer-Aided Diagnosis and Therapy
The chapters in this Medical Image Analysis and Informatics book represent some of the latest developments in the fields related to medical image analysis, medical image informatics, CBIR, and CAD. They have been prepared by leading researchers in related areas around the world.
Unlike other books in related areas, the author has chosen not to limit the applications covered by the chapters to imaging of certain parts of the body (such as the brain, the heart, or the breast) or certain diseases (such as stroke, coronary artery disease, or cancer). Instead, the range of applications is from head to toe, or craniocaudal, to use an imaging term. Several different medical imaging modalities and techniques related to CAD and image informatics are included.
The chapters should appeal to biomedical researchers, medical practitioners, neuroscientists, ophthalmologists,
dentists, radiologists, oncologists, cardiologists, orthopedic specialists, gastroenterologists, pathologists, computer scientists, medical physicists, engineers, informatics specialists, and readers interested in advanced imaging technology and informatics, and assist them in learning about a broad range of latest developments and applications in related areas.
In Chapter 1 of Medical Image Analysis and Informatics present an approach for segmentation and characterization of white matter lesions in fluid-attenuated inversion recovery (FLAIR) MR images. They describe the rationale for the use of the FLAIR modality, as well as the problem of noise in MRI and its effect on reliable segmentation. Hatanaka and Fujita present, in Chapter 2, several methods for CAD of multiple diseases via analysis of retinal fundus images.
Their methods serve the purposes of segmentation of blood vessels and measurement of vessel diameter, as well as detection of hemorrhages, microaneurysms, large cupping in the optic disc, and nerve fiber layer defects.
Why Use CAD?
At the outset, it is essential to recognize the need for application of computers for analysis of medical images. Radiologists and other medical professionals are highly trained specialists. Why, when, and for what would they need the assistance of computers? Medical images are voluminous and bear intricate details. More often than not, ordinary cases in a clinical setup or circumstances within a given model overwhelmingly outnumber unusual facts or details. Regardless of the level of expertise and experience of a medical specialist, visual analysis of medical images is prone to several types of errors.
The typical steps of a CAD system are as follows:
1. Preprocessing the given image for further analysis
2. Detection and segmentation of ROIs
3. Extraction of measures or features for quantitative analysis
4. Selection of an optimal set of features
5. Training of classifiers and development of decision rules
6. Pattern classification and diagnostic decision making