Leveraging Adaptive Color Augmentation in Convolutional Neural Networks for Deep Skin Lesion Segmentation

Abstract

Fully automatic detection of skin lesions in dermatoscopic images can facilitate early diagnosis and repression of malignant melanoma and non-melanoma skin cancer. Although convolutional neural networks are a powerful solution, they are limited by the illumination spectrum of annotated dermatoscopic screening images, where color is an important discriminative feature. In this paper, we propose an adaptive color augmentation technique to amplify data expression and model performance, while regulating color difference and saturation to minimize the risks of using synthetic data. Through deep visualization, we qualitatively identify and verify the semantic structural features learned by the network for discriminating skin lesions against normal skin tissue. The overall system achieves a Dice Ratio of 0.891 with 0.943 sensitivity and 0.932 specificity on the ISIC 2018 Testing Set for segmentation.

arXiv commentAccepted to 2020 17th IEEE International Symposium on Biomedical Imaging (ISBI) [oral presentation]

References in corpus (7)

• The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions
• Understanding Neural Networks Through Deep Visualization
• Skin Lesion Analysis Toward Melanoma Detection 2018: A Challenge Hosted by the International Skin Imaging Collaboration (ISIC)
• Automatic skin lesion segmentation with fully convolutional-deconvolutional networks
• Rethinking ImageNet Pre-training
• Data-Driven Color Augmentation Techniques for Deep Skin Image Analysis
• Deep Convolutional Encoder-Decoders with Aggregated Multi-Resolution Skip Connections for Skin Lesion Segmentation

Cited by in corpus (1)

• MFSNet: A Multi Focus Segmentation Network for Skin Lesion Segmentation