Multi-atlas segmentation: Label propagation and fusion based approach

Abstract

To understand the profundity of the subject in importance, revision of previous work has always played a vigorous role in developing interest and curiosity. For morphological assessment and measurement of quantitative parameters of biomedical structures, segmentation is done. Number of segmentation techniques has been widely used in the field of image processing since four decades. However, the problems related to segmentation still remain candid providing no optimum solution. Segmentation process is always considered as difficult due to a variation in medical images, image resolution, pixel intensity, signal variability, noise, and other artifacts. From the previous study, multi-atlas segmentation (MAS) techniques have proven to be a flexible and robust approach for medical images. Multi-atlas segmentation works in two steps: first propagation of the manually labeled images to the target image and then combining the transfer images to get the best segmentation result. Label propagation and label fusion using multiple atlases have made multi-atlas segmentation approach as forefront of segmentation research. This survey paper provides a snapshot of the current progress in the field of segmentation, registration, and label propagation.