In this paper, we present an Internet-based melanoma screening system that newly supports acral volar lesions. A half of Asian melanomas are from these areas and they show completely different appearance from other lesions. Our screening system is accessible from all over the world and diagnoses dermoscopy images within 3-5 sec based on a neural network classifier for non-acral lesions or newly integrated linear classifier for acral volar lesions. Our system achieves a sensitivity of 85.9% and a specificity of 86.0% on a set of 1258 non-acral dermoscopy images and a sensitivity of 93.3% and a specificity of 91.1% on a set of 199 acral volar dermoscopy images using a leave-one-out cross-validation.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.1109/IEMBS.2008.4650375 | DOI Listing |
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!