Few-shot Learning with Deep Triplet Networks for Brain Imaging Modality Recognition

TL;DR

This paper introduces a few-shot learning approach using Deep Triplet Networks to improve brain imaging modality recognition with limited data, outperforming traditional CNNs especially with noisy samples.

Contribution

The work presents a novel few-shot learning model based on Deep Triplet Networks tailored for brain imaging modality recognition with scarce training data.

Findings

The proposed model outperforms CNN classifiers with limited samples.

It maintains robustness against noisy and out-of-sample data.

Initial methods for incorporating uncertainty are discussed.

Abstract

Image modality recognition is essential for efficient imaging workflows in current clinical environments, where multiple imaging modalities are used to better comprehend complex diseases. Emerging biomarkers from novel, rare modalities are being developed to aid in such understanding, however the availability of these images is often limited. This scenario raises the necessity of recognising new imaging modalities without them being collected and annotated in large amounts. In this work, we present a few-shot learning model for limited training examples based on Deep Triplet Networks. We show that the proposed model is more accurate in distinguishing different modalities than a traditional Convolutional Neural Network classifier when limited samples are available. Furthermore, we evaluate the performance of both classifiers when presented with noisy samples and provide an initial inspection of how the proposed model can incorporate measures of uncertainty to be more robust against out-of-sample examples.