Fine-tuning pre-trained convolutional neural networks (CNNs) has been shown to work well for skin lesion classification . Skin cancer is among the most common cancer types . Dermoscopic image analysis improves the diagnostic accuracy for detection of malignant melanoma and other pigmented skin lesions when compared to unaided visual inspection… Hence, computer-based methods to support medical experts in the diagnostic procedure are of great interest . Using very small images (of size 64×64 pixels) degrades the classification performance, while images of size 128×128 pixels and above support good performance with larger image sizes leading to slightly improved classification . We further propose a novel fusion approach based on a three-level ensemble strategy that exploits multiple fine-tuned networks trained with dermoscopic images at various sizes . When applied on the ISIC 2017 skin lesions classification challenge, our fusion approach yields an area under the ISICS 2017 skin lesions classification, respectively, outperforming state-of-the-art algorithms, respectively,. respectively, outperforming the state-to-be-designated algorithms.  ‘Fusion approach’s ‘area under the receiver operating characteristic curve of 89.2% and 96.

Links: PDF - Abstract

Code :


Keywords : classification - skin - images - fusion - lesions -

Leave a Reply

Your email address will not be published. Required fields are marked *


Enjoy this blog? Please spread the word :)