Skin Lesion Classification Using CNNs with Patch-Based Attention and Diagnosis-Guided Loss Weighting
Abstract
Objective: This work 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 which 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 7%. Class balancing significantly improves the mean sensitivity and we show that our diagnosis-guided loss weighting method improves the mean sensitivity by 3% 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.
Nils Gessert
Research Scientist
Thilo Sentker
Research Scientist
Frederic Madesta
PhD Student
Rüdiger Schmitz
PhD Student
Helge Kniep
PhD Student
Ivo Matteo Baltruschat
Co-Founder / Research Scientist
-Ivo Matteo Baltruschat studied Information and Electrical Engineering at the University of Applied Sciences, Hamburg between 2010 and 2014. In 2016, he finished his Master of Science in medical engineering science at the Universität zu Lübeck with his thesis on Deep Learning for Advanced Medical Applications. Currently, he is a PhD student in the group of Tobias Knopp at the Institute for Biomedical Imaging, Hamburg University of Technology. His research covers the automatic analysis of medical x-ray images using machine learning methods. Our objective is to apply deep learning methods to high-resolution medical x-ray images. In the long-term, we are pursuing the goal to employ deep learning methods in order to find patterns between medical reports and medical x-ray images that will help to push the state-of-the-art in computer-aided diagnosis for chest x-ray images.
