ABSTRACT
Today numerous imaging systems are utilized in medical treatment. There are a variety of imaging techniques for the diagnosis of brain tumors, lung nodules and liver tumors and also treatment of various illnesses. Magnetic resonance imaging (MRI) is as most common as computerized tomography (CT). The techniques introduced for disease diagnosis purposes through medical images based on image processing, which result in reduced need for the workforce and, nearly, elimination of human error. More important, it is accompanied by reduced treatment expenses that in turn play a significant role in medical profession. Medical image segmentation refers to the identification, labeling, and classification of pixels in a way that the region of interest has semantic meaning. During segmentation, images are partitioned into separate segments, each of which has common features and characteristics, including similar brightness distribution. Precise recognition of tissue boundary, more specifically differentiation of tumor tissues from the healthy ones, is of significance in medical imaging. The random tumor position and shape, and poor contrast of medical images, however, are among the main challenges faced by the physicians in the diagnosis of tumors at MRI and CT scan images. This chapter aims to provide a comprehensive review of the most recent segmentation techniques of various medical images. Accordingly, a variety of medical image segmentation techniques are introduced first; and then, previous studies and literature on the diagnosis of lung, brain, and liver tumors are reviewed.
