Melanoma is an aggressive skin cancer prevalent among the Caucasian population. The site of melanoma is gender dependent: for women the prevalent site is the legs. Success of the treatment strongly correlates with early dysplastic nevus excision. Diagnosis relies on full-body skin examination and baseline photography to locate evolving, and therefore suspicious, moles.

We designed the first semi-automated legs scanner which uses one camera and laser range finder to capture images and measurements respectively. The scanner circulates around the leg, along its full length. From the collected data, a 3D leg model with moles is created. The moles are registered over temporal scans for comparison and detection of moles' evolution.

Our algorithm was tested on simulated leg model that changes over time. The average accuracy of mole registration from two temporal scans at optimal scanner settings was 94.87% with standard deviation of 6.10.