cRedAnno+: Annotation Exploitation in Self-Explanatory Lung Nodule Diagnosis

TL;DR

This paper enhances self-explanatory lung nodule diagnosis models by introducing an annotation exploitation mechanism that significantly reduces annotation needs while improving accuracy and robustness, especially under scarce annotation conditions.

Contribution

It proposes a semi-supervised active learning approach with sparse seeding and training quenching to improve annotation efficiency and model robustness in lung nodule diagnosis.

Findings

Achieves comparable or higher accuracy with 10x fewer annotations.

Demonstrates robustness and attribute prediction improvements with only 1% annotations.

Provides open-source code for reproducibility.

Abstract

Recently, attempts have been made to reduce annotation requirements in feature-based self-explanatory models for lung nodule diagnosis. As a representative, cRedAnno achieves competitive performance with considerably reduced annotation needs by introducing self-supervised contrastive learning to do unsupervised feature extraction. However, it exhibits unstable performance under scarce annotation conditions. To improve the accuracy and robustness of cRedAnno, we propose an annotation exploitation mechanism by conducting semi-supervised active learning with sparse seeding and training quenching in the learned semantically meaningful reasoning space to jointly utilise the extracted features, annotations, and unlabelled data. The proposed approach achieves comparable or even higher malignancy prediction accuracy with 10x fewer annotations, meanwhile showing better robustness and nodule attribute prediction accuracy under the condition of 1% annotations. Our complete code is open-source available: https://github.com/diku-dk/credanno.