Bag of Tricks for Long-Tailed Multi-Label Classification on Chest X-Rays
This work addresses the coupled challenges of class imbalance and label co-occurrence in chest X-ray diagnosis, though it appears incremental as it combines existing methods rather than introducing fundamentally new approaches.
The authors tackled the problem of long-tailed multi-label classification on chest X-rays by integrating multiple advanced techniques including data augmentation, feature extraction, and loss reweighting, achieving 0.349 mAP on a competition test set and ranking in the top five.
Clinical classification of chest radiography is particularly challenging for standard machine learning algorithms due to its inherent long-tailed and multi-label nature. However, few attempts take into account the coupled challenges posed by both the class imbalance and label co-occurrence, which hinders their value to boost the diagnosis on chest X-rays (CXRs) in the real-world scenarios. Besides, with the prevalence of pretraining techniques, how to incorporate these new paradigms into the current framework lacks of the systematical study. This technical report presents a brief description of our solution in the ICCV CVAMD 2023 CXR-LT Competition. We empirically explored the effectiveness for CXR diagnosis with the integration of several advanced designs about data augmentation, feature extractor, classifier design, loss function reweighting, exogenous data replenishment, etc. In addition, we improve the performance through simple test-time data augmentation and ensemble. Our framework finally achieves 0.349 mAP on the competition test set, ranking in the top five.