Medical Image Processing

Introduction. The discovery of x rays in 1895 by Professor Wilhelm Conrad Röntgen led to a transformation of medicine and science. In medicine, x rays provided a noninvasive way to visualize the internal parts of the body. A beam of radiation passing through the body is absorbed and scattered by tissue and bone structures in the path of the beam to varying extents depending on their composition and the energy level of the beam.

The resulting absorption and scatter patterns are captured by a film that is exposed during imaging to produce an image of the tissues and bone structures. By using varying amounts of energy levels of different sources of radiant energy, radiographic images can be produced for different tissues, organs and bone structures.

The simple planar x-ray imaging, the main radiologic imaging method used for most of the last century, produced high-quality analog two-dimensional (2-D) projected images of three-dimensional (3-D) organs. Over the last few decades, increasingly sophisticated methods of diagnosis have been made possible by using different types of radiant energy, including X rays, gamma rays, radio waves, and ultrasound waves.

The introduction of the first x-ray computed tomography (x-ray CT) scanner in the early 1970s totally changed the medical imaging landscape. The CT scanner uses instrumentation and computer technology for image reconstruction to produce images of cross sections of the human body. With the clinical experience accumulated over the years and the establishment of its usefulness, the CT scanner became very popular.

The exceptional multidimensional digital images of internal anatomy produced by medical imaging technology can be processed and manipulated using a computer to visualize subtle or hidden features that are not easily visible. Medical image analysis and processing algorithms for enhancing the features of interest for easy analysis and quantification are rapidly expanding the role of medical imaging beyond noninvasive examination to a tool for aiding surgical planning and intraoperative navigation. Extracting information about the shape details of anatomical structures, for example, enables careful preoperative planning of surgical procedures.

In medical image analysis, the goal is to accurately and efficiently extract information from medical images to support a range of medical investigations and clinical activities from diagnosis to surgery. The extraction of information about anatomical structures from medical images is fairly complex. This information has led to many algorithms that have been specifically proposed for biomedical applications, such as the quantification of tissue volumes, diagnosis, localization of pathology, study of anatomical structures, treatment planning, partial volume correction of functional imaging data, and computer-integrated surgery.

Technological advances in medical imaging modalities have provided doctors with significant capabilities for accurate noninvasive examination. In modern medical imaging systems, the ability to effectively process and analyze medical images to accurately extract, quantify, and interpret information to achieve an understanding of the internal structures being imaged is critical in order to support a spectrum of medical activities from diagnosis, to radiotherapy, to surgery.

Advances in computer technology and microelectronics have made available significant computational power in small desktop computers. This capability has spurred the development of software-based biomedical image analysis methods such as image enhancement and restoration, segmentation of internal structures and features of interest, image classification, and quantification.

In the next section, different medical imaging modalities are discussed before the most common methods for medical image processing and analysis are presented. An exhaustive survey of such methods can be found in Refs. 9 and 10.