Multi-instance Learning as Downstream Task of Self-Supervised Learning-based Pre-trained Model
This addresses a domain-specific problem for medical imaging researchers by improving classification in brain hematoma CT, though it is incremental as it adapts existing self-supervised learning methods to a specific bottleneck.
The paper tackled the difficulty of deep multi-instance learning when the number of instances per bag increases to 256 in brain hematoma CT images, by using a self-supervised pre-trained model as a downstream task, resulting in accuracy improvements of 5% to 13% and F1 measure gains of 40% to 55% for hypodensity marker classification.
In deep multi-instance learning, the number of applicable instances depends on the data set. In histopathology images, deep learning multi-instance learners usually assume there are hundreds to thousands instances in a bag. However, when the number of instances in a bag increases to 256 in brain hematoma CT, learning becomes extremely difficult. In this paper, we address this drawback. To overcome this problem, we propose using a pre-trained model with self-supervised learning for the multi-instance learner as a downstream task. With this method, even when the original target task suffers from the spurious correlation problem, we show improvements of 5% to 13% in accuracy and 40% to 55% in the F1 measure for the hypodensity marker classification of brain hematoma CT.