Abstract:Pneumothorax is one of the common diseases of the lungs and can be fatal in severe cases. With the development of deep learning, the convolutional neural network-based methods have been well used in the field of medical images. However, the chest X-ray images themselves are very complex, it is difficult for current algorithms to achieve pneumothorax segmentation well. Therefore, to address the problem that current deep learning-based segmentation methods cannot meet the demand of existing pneumothorax image segmentation tasks, this paper presents a novel spatially adaptive weighted fusion Signed Distance Map (SAWFSDM) network segmentation model. The proposed model utilizes the rich boundary information in the Signed Distance Map (SDM) to improve the segmentation results by learning the binary mask and SDM symbolic distance map in parallel through a multi-task learning strategy, which provides more supervised information to the training of the model through multiple labels. In addition, the boundary information learned from the SDM branch is fused into the binary mask branch during the decoding process, which strengthens the ability of model feature extraction and makes the model learn more accurate contour shapes. The segmentation model proposed in this paper has been verified by a large number of experiments on the SIIM-ACR pneumothorax segmentation dataset, and the results show that the SAWFSDM model can effectively sense the boundaries of the pneumothorax and improve the accuracy of pneumothorax segmentation, which has significant advantages compared with other advanced pneumothorax segmentation methods.

Hierarchical Lung Field Segmentation with Joint Shape and Appearance Sparse Learning.

Lung Field Segmentation Using Weighted Sparse Shape Composition with Robust Initialization

Semi-Supervised Learning for Semantic Segmentation of Emphysema With Partial Annotations

A Generic Approach to Lung Field Segmentation from Chest Radiographs using Deep Space and Shape Learning

Automatic Pathological Lung Segmentation in Low-Dose CT Image Using Eigenspace Sparse Shape Composition

Multi-stage learning for robust lung segmentation in challenging CT volumes.

Simultaneous Lung Field Detection and Segmentation for Pediatric Chest Radiographs

Lung region segmentation based on multi-resolution Active Shape Model

Towards Robust Lung Segmentation in Chest Radiographs with Deep Learning

[Non-local attention and multi-task learning based lung segmentation in chest X-ray]

TAE-Seg: Generalized Lung Segmentation Via Tilewise AutoEncoder Enhanced Network.

XLSor: A Robust and Accurate Lung Segmentor on Chest X-Rays Using Criss-Cross Attention and Customized Radiorealistic Abnormalities Generation

Segmentation of Lung Region for Chest X-Ray Images Based on Medical Registration and ASM

A Hierarchical Local Region-Based Sparse Shape Composition for Liver Segmentation in CT Scans.

Deep Belief Network and Closed Polygonal Line for Lung Segmentation in Chest Radiographs

H-SegNet: hybrid segmentation network for lung segmentation in chest radiographs using mask region-based convolutional neural network and adaptive closed polyline searching method

A full-scale lung image segmentation algorithm based on hybrid skip connection and attention mechanism

Automatic Segmentation of Pneumothorax in Chest Radiographs Based on Dual-Task Interactive Learning Method

Hybrid Automatic Lung Segmentation on Chest CT Scans

Weakly Supervised Segmentation Framework with Uncertainty: A Study on Pneumothorax Segmentation in Chest X-ray.

Rib segmentation algorithm for X-ray image based on unpaired sample augmentation and multi-scale network