Differentiable Topology-Preserved Distance Transform for Pulmonary Airway Segmentation
Not on the map. Why?
Abstract (arXiv)
Detailed pulmonary airway segmentation is a clinically important task for endobronchial intervention and treatment of peripheral located lung cancer lesions. Convolutional Neural Networks (CNNs) are promising tools for medical image analysis but have been performing poorly for cases when existing a significant imbalanced feature distribution, which is true for the airway data as the trachea and principal bronchi dominate most of the voxels whereas the lobar bronchi and distal segmental bronchi occupy a small proportion. In this paper, we propose a Differentiable Topology-Preserved Distance Transform (DTPDT) framework to improve the performance of airway segmentation. A Topology-Preserved Surrogate (TPS) learning strategy is first proposed to balance the training progress within-class distribution. Furthermore, a Convolutional Distance Transform (CDT) is designed to identify the breakage phenomenon with superior sensitivity and minimize the variation of the distance map between the predictionand ground-truth. The proposed method is validated with the publically available reference airway segmentation datasets. The detected rate of branch and length on public EXACT'09 and BAS datasets are 82.1%/79.6% and 96.5%/91.5% respectively, demonstrating the reliability and efficiency of the method in terms of improving the topology completeness of the segmentation performance while maintaining the overall topology accuracy.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 20 of its 37 references that are in the map. First posted on arXiv 17 Sep 2022.
Builds on
20 of 37 references.
References · 20
- 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,590
- 2016Categorical Reparameterization with Gumbel-Softmaxcited by 2,708
- 20163D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotationcited by 1,471
- 2019Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regressioncited by 1,361
- 2016Instance Normalization: The Missing Ingredient for Fast Stylizationcited by 1,187
- 2017A systematic study of the class imbalance problem in convolutional neural networkscited by 510
- 2020clDice -- A Novel Topology-Preserving Loss Function for Tubular Structure Segmentationcited by 286
- 2017Tversky loss function for image segmentation using 3D fully convolutional deep networkscited by 228
- 2018Boundary loss for highly unbalanced segmentationcited by 187
- 2019Deep Distance Transform for Tubular Structure Segmentation in CT Scanscited by 52
- 2019Shape-Aware Organ Segmentation by Predicting Signed Distance Mapscited by 41
- 2020Alleviating Class-wise Gradient Imbalance for Pulmonary Airway Segmentationcited by 34
8 more
- 2019AirwayNet: A Voxel-Connectivity Aware Approach for Accurate Airway Segmentation Using Convolutional Neural Networkscited by 32
- 2020Learning Tubule-Sensitive CNNs for Pulmonary Airway and Artery-Vein Segmentation in CTcited by 28
- 2019A joint 3D UNet-Graph Neural Network-based method for Airway Segmentation from chest CTscited by 27
- 2018Automatic Airway Segmentation in chest CT using Convolutional Neural Networkscited by 18
- 2022BREAK: Bronchi Reconstruction by gEodesic transformation And sKeleton embeddingcited by 16
- 2019Shape-Aware Complementary-Task Learning for Multi-Organ Segmentationcited by 15
- 2020Learning Surrogates via Deep Embeddingcited by 11
- 2022LTSP: Long-Term Slice Propagation for Accurate Airway Segmentationcited by 6
Built on by
Citing papers per year: 2023: 2.
2 papers.