# Alleviating Class-wise Gradient Imbalance for Pulmonary Airway Segmentation

arXiv: https://arxiv.org/abs/2011.11952 · HTML: https://mlsanity.com/paper/2011.11952/

## Key facts

- Cited by 35 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Nov 2020.

## Built on by (most cited first, 30 of 35)

- [3D MedDiffusion: A 3D Medical Latent Diffusion Model for Controllable and High-quality Medical Image Generation](https://mlsanity.com/paper/2412.13059/) (2024)
- [Efficient automatic segmentation for multi-level pulmonary arteries: The PARSE challenge](https://mlsanity.com/paper/2304.03708/) (2023)
- [Automatic airway segmentation from Computed Tomography using robust and efficient 3-D convolutional neural networks](https://mlsanity.com/paper/2103.16328/) (2021)
- [Fuzzy Attention Neural Network to Tackle Discontinuity in Airway Segmentation](https://mlsanity.com/paper/2209.02048/) (2022)
- [NaviAirway: a Bronchiole-sensitive Deep Learning-based Airway Segmentation Pipeline](https://mlsanity.com/paper/2203.04294/) (2022)
- [BREAK: Bronchi Reconstruction by gEodesic transformation And sKeleton embedding](https://mlsanity.com/paper/2202.00002/) (2022)
- [Human Treelike Tubular Structure Segmentation: A Comprehensive Review and Future Perspectives](https://mlsanity.com/paper/2207.11203/) (2022)
- [Hunting imaging biomarkers in pulmonary fibrosis: Benchmarks of the AIIB23 challenge](https://mlsanity.com/paper/2312.13752/) (2023)
- [BronchoTrack: Airway Lumen Tracking for Branch-Level Bronchoscopic Localization](https://mlsanity.com/paper/2402.12763/) (2024)
- [Two-stage Contextual Transformer-based Convolutional Neural Network for Airway Extraction from CT Images](https://mlsanity.com/paper/2212.07651/) (2022)
- [AeroPath: An airway segmentation benchmark dataset with challenging pathology](https://mlsanity.com/paper/2311.01138/) (2023)
- [Multi-site, Multi-domain Airway Tree Modeling (ATM'22): A Public Benchmark for Pulmonary Airway Segmentation](https://mlsanity.com/paper/2303.05745/) (2023)
- [Accurate Airway Tree Segmentation in CT Scans via Anatomy-aware Multi-class Segmentation and Topology-guided Iterative Learning](https://mlsanity.com/paper/2306.09116/) (2023)
- [Intensity Confusion Matters: An Intensity-Distance Guided Loss for Bronchus Segmentation](https://mlsanity.com/paper/2406.16150/) (2024)
- [BCNet: Bronchus Classification via Structure Guided Representation Learning](https://mlsanity.com/paper/2205.06947/) (2022)
- [Coarse-to-fine Airway Segmentation Using Multi information Fusion Network and CNN-based Region Growing](https://mlsanity.com/paper/2102.12755/) (2021)
- [Adversarial Transformer for Repairing Human Airway Segmentation](https://mlsanity.com/paper/2210.12029/) (2022)
- [Trexplorer Super: Topologically Correct Centerline Tree Tracking of Tubular Objects in CT Volumes](https://mlsanity.com/paper/2507.10881/) (2025)
- [GDDS: Pulmonary Bronchioles Segmentation with Group Deep Dense Supervision](https://mlsanity.com/paper/2303.09212/) (2023)
- [HarmonySeg: Tubular Structure Segmentation with Deep-Shallow Feature Fusion and Growth-Suppression Balanced Loss](https://mlsanity.com/paper/2504.07827/) (2025)
- [Extraction of Pulmonary Airway in CT Scans Using Deep Fully Convolutional Networks](https://mlsanity.com/paper/2208.07202/) (2022)
- [Differentiable Topology-Preserved Distance Transform for Pulmonary Airway Segmentation](https://mlsanity.com/paper/2209.08355/) (2022)
- [Open-source tool for Airway Segmentation in Computed Tomography using 2.5D Modified EfficientDet: Contribution to the ATM22 Challenge](https://mlsanity.com/paper/2209.15094/) (2022)
- [MEDPSeg: Hierarchical polymorphic multitask learning for the segmentation of ground-glass opacities, consolidation, and pulmonary structures on computed tomography](https://mlsanity.com/paper/2312.02365/) (2023)
- [BronchoCopilot: Towards Autonomous Robotic Bronchoscopy via Multimodal Reinforcement Learning](https://mlsanity.com/paper/2403.01483/) (2024)
- [Progressive Curriculum Learning with Scale-Enhanced U-Net for Continuous Airway Segmentation](https://mlsanity.com/paper/2410.18456/) (2024)
- [Make it more specific: A novel uncertainty based airway segmentation application on 3D U-Net and its variants](https://mlsanity.com/paper/2402.07403/) (2024)
- [TopoSculpt: Betti-Steered Topological Sculpting of 3D Fine-grained Tubular Shapes](https://mlsanity.com/paper/2509.03938/) (2025)
- [Airway Tree Modeling Using Dual-channel 3D UNet 3+ with Vesselness Prior](https://mlsanity.com/paper/2208.13969/) (2022)
- [A CT-Based Airway Segmentation Using U2-net Trained by the Dice Loss Function](https://mlsanity.com/paper/2209.10796/) (2022)

## Abstract (arXiv)

Automated airway segmentation is a prerequisite for pre-operative diagnosis and intra-operative navigation for pulmonary intervention. Due to the small size and scattered spatial distribution of peripheral bronchi, this is hampered by severe class imbalance between foreground and background regions, which makes it challenging for CNN-based methods to parse distal small airways. In this paper, we demonstrate that this problem is arisen by gradient erosion and dilation of the neighborhood voxels. During back-propagation, if the ratio of the foreground gradient to background gradient is small while the class imbalance is local, the foreground gradients can be eroded by their neighborhoods. This process cumulatively increases the noise information included in the gradient flow from top layers to the bottom ones, limiting the learning of small structures in CNNs. To alleviate this problem, we use group supervision and the corresponding WingsNet to provide complementary gradient flows to enhance the training of shallow layers. To further address the intra-class imbalance between large and small airways, we design a General Union loss function which obviates the impact of airway size by distance-based weights and adaptively tunes the gradient ratio based on the learning process. Extensive experiments on public datasets demonstrate that the proposed method can predict the airway structures with higher accuracy and better morphological completeness than the baselines.