Learning to Learn Unlearned Feature for Brain Tumor Segmentation
This work addresses the challenge of medical image segmentation for brain tumors, where data scarcity and disease variants like glioma and metastasis hinder large-scale dataset creation, offering a solution for efficient transfer learning in this domain-specific context.
The paper tackles the problem of brain tumor segmentation with limited annotated data by proposing a fine-tuning algorithm that combines active and meta-learning to transfer knowledge from high-grade glioma to brain metastasis, achieving balanced parameters for both domains in a few steps.
We propose a fine-tuning algorithm for brain tumor segmentation that needs only a few data samples and helps networks not to forget the original tasks. Our approach is based on active learning and meta-learning. One of the difficulties in medical image segmentation is the lack of datasets with proper annotations, because it requires doctors to tag reliable annotation and there are many variants of a disease, such as glioma and brain metastasis, which are the different types of brain tumor and have different structural features in MR images. Therefore, it is impossible to produce the large-scale medical image datasets for all types of diseases. In this paper, we show a transfer learning method from high grade glioma to brain metastasis, and demonstrate that the proposed algorithm achieves balanced parameters for both glioma and brain metastasis domains within a few steps.