Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Loss Design and Architecture Selection for Long-Tailed Multi-Label Chest X-Ray Classification

Nikhileswara Rao Sulake

TL;DR

A systematic empirical evaluation of loss functions, CNN backbone architectures and post-training strategies on the CXR-LT 2026 benchmark, comprising approximately 143K images with 30 disease labels from PadChest shows that LDAM with deferred re-weighting (LDAM-DRW) consistently outperforms standard BCE and asymmetric losses for rare class recognition.

Abstract

Long-tailed class distributions pose a significant challenge for multi-label chest X-ray (CXR) classification, where rare but clinically important findings are severely underrepresented. In this work, we present a systematic empirical evaluation of loss functions, CNN backbone architectures and post-training strategies on the CXR-LT 2026 benchmark, comprising approximately 143K images with 30 disease labels from PadChest. Our experiments demonstrate that LDAM with deferred re-weighting (LDAM-DRW) consistently outperforms standard BCE and asymmetric losses for rare class recognition. Amongst the architectures evaluated, ConvNeXt-Large achieves the best single-model performance with 0.5220 mAP and 0.3765 F1 on our development set, whilst classifier re-training and test-time augmentation further improve ranking metrics. On the official test leaderboard, our submission achieved 0.3950 mAP, ranking 5th amongst all 68 participating teams with total of 1528 submissions. We provide a candid analysis of the development-to-test performance gap and discuss practical insights for handling class imbalance in clinical imaging settings. Code is available at https://github.com/Nikhil-Rao20/Long_Tail.

Loss Design and Architecture Selection for Long-Tailed Multi-Label Chest X-Ray Classification

TL;DR

A systematic empirical evaluation of loss functions, CNN backbone architectures and post-training strategies on the CXR-LT 2026 benchmark, comprising approximately 143K images with 30 disease labels from PadChest shows that LDAM with deferred re-weighting (LDAM-DRW) consistently outperforms standard BCE and asymmetric losses for rare class recognition.

Abstract

Long-tailed class distributions pose a significant challenge for multi-label chest X-ray (CXR) classification, where rare but clinically important findings are severely underrepresented. In this work, we present a systematic empirical evaluation of loss functions, CNN backbone architectures and post-training strategies on the CXR-LT 2026 benchmark, comprising approximately 143K images with 30 disease labels from PadChest. Our experiments demonstrate that LDAM with deferred re-weighting (LDAM-DRW) consistently outperforms standard BCE and asymmetric losses for rare class recognition. Amongst the architectures evaluated, ConvNeXt-Large achieves the best single-model performance with 0.5220 mAP and 0.3765 F1 on our development set, whilst classifier re-training and test-time augmentation further improve ranking metrics. On the official test leaderboard, our submission achieved 0.3950 mAP, ranking 5th amongst all 68 participating teams with total of 1528 submissions. We provide a candid analysis of the development-to-test performance gap and discuss practical insights for handling class imbalance in clinical imaging settings. Code is available at https://github.com/Nikhil-Rao20/Long_Tail.
Paper Structure (12 sections, 2 equations, 1 figure, 2 tables)

This paper contains 12 sections, 2 equations, 1 figure, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Problem Formulation
  4. Loss Functions for Long-Tailed Data
  5. Network Architectures
  6. Post-Training Strategies: cRT, TTA and Ensemble
  7. Experiments
  8. Dataset and Preprocessing
  9. Training Details
  10. Results and Discussion
  11. Conclusion
  12. Acknowledgments

Figures (1)

  • Figure 1: Class-activation maps overlaid on CXR test images (predicted label and probability shown). The model localizes many findings correctly but probability calibration and thresholding cause instance-level misses. Rows correspond to different findings (kyphosis, hernia, azygos lobe).