Skin Lesion Classification Using CNNs With Patch-Based Attention and Diagnosis-Guided Loss Weighting

IEEE Trans Biomed Eng. 2020 Feb;67(2):495-503. doi: 10.1109/TBME.2019.2915839. Epub 2019 May 9.

Abstract

Objective: This paper addresses two key problems of skin lesion classification. The first problem is the effective use of high-resolution images with pretrained standard architectures for image classification. The second problem is the high-class imbalance encountered in real-world multi-class datasets.

Methods: To use high-resolution images, we propose a novel patch-based attention architecture that provides global context between small, high-resolution patches. We modify three pretrained architectures and study the performance of patch-based attention. To counter class imbalance problems, we compare oversampling, balanced batch sampling, and class-specific loss weighting. Additionally, we propose a novel diagnosis-guided loss weighting method that takes the method used for ground-truth annotation into account.

Results: Our patch-based attention mechanism outperforms previous methods and improves the mean sensitivity by [Formula: see text]. Class balancing significantly improves the mean sensitivity and we show that our diagnosis-guided loss weighting method improves the mean sensitivity by [Formula: see text] over normal loss balancing.

Conclusion: The novel patch-based attention mechanism can be integrated into pretrained architectures and provides global context between local patches while outperforming other patch-based methods. Hence, pretrained architectures can be readily used with high-resolution images without downsampling. The new diagnosis-guided loss weighting method outperforms other methods and allows for effective training when facing class imbalance.

Significance: The proposed methods improve automatic skin lesion classification. They can be extended to other clinical applications where high-resolution image data and class imbalance are relevant.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Databases, Factual
  • Deep Learning*
  • Dermoscopy
  • Humans
  • Image Interpretation, Computer-Assisted / methods*
  • Skin / diagnostic imaging
  • Skin Neoplasms / classification
  • Skin Neoplasms / diagnostic imaging*
  • Skin Neoplasms / pathology