Skin Lesion Diagnosis using Ensembles, Unscaled Multi-Crop Evaluation and Loss Weighting
This work addresses automated skin lesion diagnosis for medical imaging, but it is incremental as it builds on existing deep learning models and techniques.
The paper tackled skin lesion diagnosis by using an ensemble of convolutional neural networks with loss weighting and unscaled multi-crop evaluation to address class imbalance, achieving second place in the ISIC 2018 challenge as the best method using only publicly available data.
In this paper we present the methods of our submission to the ISIC 2018 challenge for skin lesion diagnosis (Task 3). The dataset consists of 10000 images with seven image-level classes to be distinguished by an automated algorithm. We employ an ensemble of convolutional neural networks for this task. In particular, we fine-tune pretrained state-of-the-art deep learning models such as Densenet, SENet and ResNeXt. We identify heavy class imbalance as a key problem for this challenge and consider multiple balancing approaches such as loss weighting and balanced batch sampling. Another important feature of our pipeline is the use of a vast amount of unscaled crops for evaluation. Last, we consider meta learning approaches for the final predictions. Our team placed second at the challenge while being the best approach using only publicly available data.