M-VAAL: Multimodal Variational Adversarial Active Learning for Downstream Medical Image Analysis Tasks

TL;DR

This paper introduces M-VAAL, a multimodal active learning approach that leverages auxiliary information from multiple modalities to improve data efficiency in medical image analysis tasks, reducing annotation costs.

Contribution

The work presents a novel multimodal variational adversarial active learning method that incorporates auxiliary modalities to enhance sample selection in medical imaging, addressing limitations of previous task-specific approaches.

Findings

Improved performance in brain tumor segmentation and classification.

Effective data reduction with limited annotations.

Enhanced robustness through multimodal auxiliary information.

Abstract

Acquiring properly annotated data is expensive in the medical field as it requires experts, time-consuming protocols, and rigorous validation. Active learning attempts to minimize the need for large annotated samples by actively sampling the most informative examples for annotation. These examples contribute significantly to improving the performance of supervised machine learning models, and thus, active learning can play an essential role in selecting the most appropriate information in deep learning-based diagnosis, clinical assessments, and treatment planning. Although some existing works have proposed methods for sampling the best examples for annotation in medical image analysis, they are not task-agnostic and do not use multimodal auxiliary information in the sampler, which has the potential to increase robustness. Therefore, in this work, we propose a Multimodal Variational Adversarial Active Learning (M-VAAL) method that uses auxiliary information from additional modalities to enhance the active sampling. We applied our method to two datasets: i) brain tumor segmentation and multi-label classification using the BraTS2018 dataset, and ii) chest X-ray image classification using the COVID-QU-Ex dataset. Our results show a promising direction toward data-efficient learning under limited annotations.