On the generalization capabilities of FSL methods through domain adaptation: a case study in endoscopic kidney stone image classification
This work addresses domain adaptation for medical imaging, specifically endoscopic kidney stone classification, offering a solution to improve generalization across different acquisition conditions, though it is incremental as it applies existing meta-learning techniques to a new domain.
The paper tackles the problem of poor generalization of deep learning methods across domains due to data distribution shifts by proposing a meta-learning based few-shot learning approach, achieving accuracies of 74.38% and 88.52% in 5-way 5-shot and 5-way 20-shot settings on endoscopic kidney stone image datasets, compared to 45% for traditional methods.
Deep learning has shown great promise in diverse areas of computer vision, such as image classification, object detection and semantic segmentation, among many others. However, as it has been repeatedly demonstrated, deep learning methods trained on a dataset do not generalize well to datasets from other domains or even to similar datasets, due to data distribution shifts. In this work, we propose the use of a meta-learning based few-shot learning approach to alleviate these problems. In order to demonstrate its efficacy, we use two datasets of kidney stones samples acquired with different endoscopes and different acquisition conditions. The results show how such methods are indeed capable of handling domain-shifts by attaining an accuracy of 74.38% and 88.52% in the 5-way 5-shot and 5-way 20-shot settings respectively. Instead, in the same dataset, traditional Deep Learning (DL) methods attain only an accuracy of 45%.