Fig. 1.

Semi-automatic edge-based (SAEB) algorithm workflow. The algorithm requires the intervention of an operator to insert the central point of the lesion (a) through an interface created in MATLAB, the operator can view both the PET and the CT images separately. As a second step, edge-preserving filters are applied to the image to emphasize the edge of the lesion (b) and, after that, and a level-set is used which is a shape that evolves iteratively over the image (c). The level-set acts both on the original image of the lesion and on the filtered image in which the contours are highlighted. The iteration 0 of the level set, which would be the initialization, is the center of the lesion indicated by the operator. The final outline of the lesion is achieved at the end of the process