- Applications: Image Alignment
Ayman S El-Baz, Georgy Gimel’farb, Jasjit S. Suri · 2015
Image alignment, which is called also image registration, mapping or matching, superposes two or more images by establishing their pixel-to-pixel or voxel-to-voxel correspondences and bringing the corresponding pixels or voxels into coincidence. One of the most common goals is to closely superpose the same or similar objects-of-interest in two or more images taken, e.g., at different times, from different viewpoints, and/or with different modalities (sensors). The alignment is a crucial step in many applications of medical image analysis for clinical diagnostics, therapy planning, and more complete descriptions of anatomical objects that require an accurate fusion of different imaging modalities, like CT and MRI; detection of temporal changes in a sequence of images; integration of separate images, e.g., to form a 3D (volumetric) image from separate 2D slices, and so forth. Co-registered and fused MRI, CT, PET, SPECT, US, and other medical images not only provide more detailed information about a patient, but also ensure monitoring a disease, e.g., a tumor growth, verifying a treatment, and comparing patient’s data to anatomical atlases or other patients.