Liver segmentation for hepatic lesions detection and characterisation

The detection and characterisation of hepatic lesions is fundamental in clinical practice, from the diagnosis stages to the evolution of the therapeutic response. Hepatic magnetic resonance is a usual pratice in the localization and quantification of lesions. Automatic segmentation of the liver is illustrated in T <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> weighted images. This task is necessary for detecting the lesions. The proposed liver segmentation is based on 3D anisotropic diffusion processing without any control parameter. Combinations of edge detection techniques, histogram analysis, morphological post-processing and evolution of an active contour have been applied to the liver segmentation. The active contour evolution is based on the minimization of variance in luminance between the liver and its closest neighbourhood.