TY - JOUR T1 - Lesion Attributes Segmentation for Melanoma Detection with Deep Learning JF - bioRxiv DO - 10.1101/381855 SP - 381855 AU - Eric Z. Chen AU - Xu Dong AU - Junyan Wu AU - Hongda Jiang AU - Xiaoxiao Li AU - Ruichen Rong Y1 - 2018/01/01 UR - http://biorxiv.org/content/early/2018/09/10/381855.abstract N2 - Melanoma is the most deadly form of skin cancer world-wide. Many efforts have been made for early detection of melanoma. The International Skin Imaging Collaboration (ISIC) hosted the 2018 Challenges to improve the diagnosis of melanoma based on dermoscopic images. In this paper, we describe our solution for the task 2 of ISIC 2018 Challenges. We present a multi-task U-Net model to automatically detect lesion attributes of melanoma. Our multi-task U-Net deep learning model achieves a Jaccard index of 0.433 on official test data, which ranks the 5th place on the final leaderboard. ER -