Saved in:
Bibliographic Details
Main Authors: Khuong, Duy Hoang, Huu, Duy Nguyen, Viet, Ngu Huynh Cong
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.02292
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915978570366976
author Khuong, Duy Hoang
Huu, Duy Nguyen
Viet, Ngu Huynh Cong
author_facet Khuong, Duy Hoang
Huu, Duy Nguyen
Viet, Ngu Huynh Cong
contents Chest X-ray classification suffers from severe class imbalance where gradient updates bias toward majority classes, causing feature drift and poor performance on rare but critical pathologies. We propose a Momentum-Anchored Multi-Scale Fusion Network that uses exponential moving averages (EMA) as a temporal anchoring mechanism to stabilize feature representations under long-tailed distributions. Our approach applies selective momentum updates to the final expansion block of an EfficientNet backbone, creating a slowly-evolving reference branch that resists gradient-induced drift while preserving discriminative patterns for minority classes. Combined with multi-scale spatial fusion ($1\times 1$, $3 \times 3$, $5 \times 5$ convolutions), this anchoring strategy maintains representational stability throughout training. On ChestX-ray14, our method achieves 0.8682 average AUC, outperforming state-of-the-art approaches and showing particular improvements on rare pathologies like Hernia (0.9470) and Pneumonia (0.8165). The results demonstrate that momentum anchoring effectively counters feature instability in long-tailed medical image classification.
format Preprint
id arxiv_https___arxiv_org_abs_2605_02292
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Momentum-Anchored Multi-Scale Fusion Model for Long-Tailed Chest X-Ray Classification
Khuong, Duy Hoang
Huu, Duy Nguyen
Viet, Ngu Huynh Cong
Computer Vision and Pattern Recognition
Chest X-ray classification suffers from severe class imbalance where gradient updates bias toward majority classes, causing feature drift and poor performance on rare but critical pathologies. We propose a Momentum-Anchored Multi-Scale Fusion Network that uses exponential moving averages (EMA) as a temporal anchoring mechanism to stabilize feature representations under long-tailed distributions. Our approach applies selective momentum updates to the final expansion block of an EfficientNet backbone, creating a slowly-evolving reference branch that resists gradient-induced drift while preserving discriminative patterns for minority classes. Combined with multi-scale spatial fusion ($1\times 1$, $3 \times 3$, $5 \times 5$ convolutions), this anchoring strategy maintains representational stability throughout training. On ChestX-ray14, our method achieves 0.8682 average AUC, outperforming state-of-the-art approaches and showing particular improvements on rare pathologies like Hernia (0.9470) and Pneumonia (0.8165). The results demonstrate that momentum anchoring effectively counters feature instability in long-tailed medical image classification.
title Momentum-Anchored Multi-Scale Fusion Model for Long-Tailed Chest X-Ray Classification
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.02292