Segmentation of pulmonary nodule images using 1-norm minimization

Total variation minimization (in the 1-norm) has edge preserving and enhanc ing properties which make it suitable for image segmentation. We present Im age Simplification, a new formulation and algorithm for image segmentation. We illustrate the edge enhancing properties of 1-norm total variation mini mization in a discrete setting by giving exact solutions to the problem for piecewise constant functions in the presence of noise. In this case, edges can be exactly recovered if the noise is sufficiently small. After optimiz ation, segmentation is completed using edge detection. We find that our ima ge segmentation approach yields good results when applied to the segmentati on of pulmonary nodules.