Background: Dermoscopy is one of the major imaging modalities used in thediagnosis of melanoma and other pigmented skin lesions. Due to the difficultyand subjectivity of human interpretation, dermoscopy image analysis has becomean important research area. One of the most important steps in dermoscopy imageanalysis is the automated detection of lesion borders. Although numerousmethods have been developed for the detection of lesion borders, very fewstudies were comprehensive in the evaluation of their results. Methods: In thispaper, we evaluate five recent border detection methods on a set of 90dermoscopy images using three sets of dermatologist-drawn borders as theground-truth. In contrast to previous work, we utilize an objective measure,the Normalized Probabilistic Rand Index, which takes into account thevariations in the ground-truth images. Conclusion: The results demonstrate thatthe differences between four of the evaluated border detection methods are infact smaller than those predicted by the commonly used XOR measure.
展开▼
