Guide to Medical Image Analysis_ Methods and Algorithms下载

The methodology presented in the first edition was considered established practice or settled science in the medical image analysis community in 2010–2011. Progress in this field is fast (as in all fields of computer science) with several developments being particularly relevant to subjects treated in this book: • Image-based guidance in the operating room is no longer restricted to the display of planning images during intervention. It is increasingly meant to aid the operator to adapt his or her intervention technique during operation. This requires reliable and intuitive analysis methods. • Segmentation and labeling of images is now mostly treated as solution of an optimization problem in the discrete (Chap. 8) or in the continuous domain (Chap. 9). Heuristic methods such as the one presented in Chap. 6 still exist in non-commercial and commercial software products, but searching for results that optimize an assumption about how the information is mapped to the data produces more predictable methods. • Deep learning gives new impulses to many areas in medical image analysis as it combines learning of features from data with the abstraction ability of multilayer perceptrons. Hence, learning strategies can be applied directly to pixels in a labeling task. It promises analysis methods that are not designed for a specific problem but can be trained from examples in this problem domain.