An image-based segmentation recommender using crowdsourcing and transfer learning for skin lesion extraction
作者：
Highlights：
• A segmentation recommender based on crowdsourcing and transfer learning.
• Two pre-trained architectures (VGG16 and ResNet50) were studied.
• The method predicts dynamically the most appropriate segmenter for any input image.
• Creation of new ground -truth dataset by using the ISIC2017 segmentation results.
• Proposed image-based method improves the performance of skin lesion segmentation.
摘要
•A segmentation recommender based on crowdsourcing and transfer learning.•Two pre-trained architectures (VGG16 and ResNet50) were studied.•The method predicts dynamically the most appropriate segmenter for any input image.•Creation of new ground -truth dataset by using the ISIC2017 segmentation results.•Proposed image-based method improves the performance of skin lesion segmentation.
论文关键词：Segmentation recommender,Deep learning,Crowdsourcing,Transfer learning,VGG16,ResNet50
论文评审过程：Received 3 May 2018, Revised 14 October 2018, Accepted 15 October 2018, Available online 15 October 2018, Version of Record 18 October 2018.
论文官网地址：https://doi.org/10.1016/j.eswa.2018.10.029
