Alleviating Class-Wise Gradient Imbalance for Pulmonary Airway Segmentation
Top Cited Papers
- 10 May 2021
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 40 (9) , 2452-2462
Abstract
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 a 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 that 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.Keywords
All Related Versions
Funding Information
- National Key Research and Development Program of China (2019YFB1311503, 2017YFC0112700)
- Committee of Science and Technology, Shanghai, China (19510711200)
- Shanghai Sailing Program (20YF1420800)
- NSFC (61661010, 61977046, 62003208)
- Shanghai Jiao Tong University (SJTU) Trans-med Awards Research (20210101)
- Science and Technology Commission of Shanghai Municipality (20DZ2220400)
This publication has 34 references indexed in Scilit:
- A hybrid method for airway segmentation and automated measurement of bronchial wall thickness on CTMedical Image Analysis, 2015
- Efficient object localization using Convolutional Networks2015
- Navigated BronchoscopyJournal of Bronchology & Interventional Pulmonology, 2014
- Automatic Airway Analysis on Multidetector Computed Tomography in Cystic FibrosisJournal of Thoracic Imaging, 2013
- Robust airway extraction based on machine learning and minimum spanning tree2013
- Extraction of Airways From CT (EXACT'09)IEEE Transactions on Medical Imaging, 2012
- WE-B-201B-02: The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI): A Completed Public Database of CT Scans for Lung Nodule AnalysisMedical Physics, 2010
- Interactive Organ Segmentation Using Graph Cuts2000
- Tissue classification based on 3D local intensity structures for volume renderingIEEE Transactions on Visualization and Computer Graphics, 2000
- Building Skeleton Models via 3-D Medial Surface Axis Thinning AlgorithmsCVGIP: Graphical Models and Image Processing, 1994