Enhancing the Generalizability of Deep Learning-Based Models for Lung Field Segmentation in Chest Radiographs Using Edge-Assisted Multiscale Feature Fusion.

Locating lung field is a critical and fundamental processing stage in the automated analysis of chest radiographs (CXRs) for pulmonary disorders. During the routine examination of CXRs, using both frontal and lateral CXRs can benefit clinical diagnosis of cardiothoracic and lung diseases. However, the accurate segmentation of lung fields on both frontal and lateral CXRs is still challenging due to the blurry boundary of the lung field on lateral CXRs and the poor generalization ability of the models. Existing deep learning-based methods focused on lung field segmentation on frontal CXRs, and the generalization ability of these methods on the different type of CXRs (e.g., pediatric CXRs) and new lung diseases (e.g., COVID-19) has not been tested. In this paper, a view identification assisted fully convolutional network (VI-FCN) is proposed for the segmentation of lung fields on frontal and lateral CXRs simultaneously. The VI-FCN consists of an FCN branch for lung field segmentation and a view identification branch for identification of the frontal and lateral CXRs and for enhancing the lung field segmentation. To improve the generalization ability of VI-FCN, six public datasets and our frontal and lateral CXRs (over 2000 CXRs) were collected for training. The segmentation of lung fields on the Japanese Society of Radiological Technology (JSRT) dataset yields mean dice similarity coefficient (DSC) of 0.979 ± 0.008, mean Jaccard index ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\Omega $ </tex-math></inline-formula> ) of 0.959 ± 0.016, and mean boundary distance (MBD) of 1.023 ± 0.487 <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">mm</i> . Besides, the VI-FCN achieves mean DSC of 0.973 ± 0.010, mean <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\Omega $ </tex-math></inline-formula> of 0.947 ± 0.018, and mean MBD of 1.923 ± 0.755 <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">mm</i> for the segmentation of lung fields on our lateral CXRs. The experiments demonstrate the superior performance of the proposed VI-FCN over most of existing state-of-the-art methods. Moreover, the proposed VI-FCN achieves promising results on untrained pediatric CXRs and COVID-19 datasets.

Lung segmentation plays an important role in computer-aided detection and diagnosis using chest radiographs (CRs). Currently, the U-Net and DeepLabv3+ convolutional neural network architectures are widely used to perform CR lung segmentation. To boost performance, ensemble methods are often used, whereby probability map outputs from several networks operating on the same input image are averaged. However, not all networks perform adequately for any specific patient image, even if the average network performance is good. To address this, we present a novel multi-network ensemble method that employs a selector network. The selector network evaluates the segmentation outputs from several networks; on a case-by-case basis, it selects which outputs are fused to form the final segmentation for that patient. Our candidate lung segmentation networks include U-Net, with five different encoder depths, and DeepLabv3+, with two different backbone networks (ResNet50 and ResNet18). Our selector network is a ResNet18 image classifier. We perform all training using the publicly available Shenzhen CR dataset. Performance testing is carried out with two independent publicly available CR datasets, namely, Montgomery County (MC) and Japanese Society of Radiological Technology (JSRT). Intersection-over-Union scores for the proposed approach are 13% higher than the standard averaging ensemble method on MC and 5% better on JSRT.

Lung Field Segmentation (LFS) is an indispensable step for detecting austere lung diseases in various computer-aided diagnosis. This paper presents a deep learning-based Convolutional Neural Network (CNN) for segmenting lung fields in chest radiographs. The proposed CNN network consists of three sets of convolutional-layer and rectified linear unit (ReLU) layer, followed by a fully connected layer. At each convolutional layer, 64 filters retrieve the representative features. Japanese Society of Radiological Technology (JSRT) dataset is used for training and validation. Test results have 98.05% average accuracy, 93.4% average overlap, 96.25% average sensitivity, and 98.80% average specificity. The obtained results are promising and better than many of the existing state-of-the-art LFS techniques.

Segmentation of the lung field is considered as the first and crucial stage in diagnosis of pulmonary diseases. In clinical practice, computer-aided systems are used to segment the lung region from chest X-ray (CXR) or CT images. The task of segmentation is challenging due to the presence of opacities or consolidation in CXR, which are typically produced by overlaps between the lung region and intense abnormalities caused by pulmonary diseases such as pneumonia, tuberculosis, or COVID-19. Recently, Convolution Neural Networks (CNNs) have been shown promising for segmentation and detection in digital images. In this paper, we propose a two-stage framework based on adapted U-Net architecture to leverage automatic lung segmentation. In the first stage, we extract CXR-patches and train a modified U-Net architecture to generate an initial segmentation of lung field. The second stage is the post-processing step, where we deploy image processing techniques to obtain a clear final segmentation. The performance of the proposed method is evaluated on a set of 138 CXR images obtained from Montgomery County’s Tuberculosis Control Program, producing an average Dice Coefficient (DC) of 94.21%, and an average Intersection-Over-Union (IoU) of 91.37%.

Accurate lung field segmentation (LFS) method is highly demanded in computer-aid diagnosis (CAD) system. However, LFS in pediatric CXR images has received few attention due to the lack of publicly available dataset and the challenges caused by their unique characteristics, such as great variations of the size, location and orientation of lungs. To fill this gap, this paper for the first time presents a simultaneous lung field detection and segmentation framework for pediatric CXR images. Our framework, called SDSLung Net, is a multi-tasking convolutional neural network architecture tailor-made for X-ray images with relatively weak appearance feature but abundant spatial rules and structural information. It is adapted from a Mask R-CNN framework [1] by incorporating a newly designed Organ Structure-Aware Encoding layer in the backbone network for more accurate spatial variation and structural representation, in parallel with a deeply supervised fully convolutional network based segmentation branch for precise lung field segmentation inside detected bounding box. Moreover, we also constructed a new and so far the largest pediatric CXR dataset with pixelwise lung field annotations. Experimental results demonstrate that our proposed SDSLung is capable of achieving significantly superior performance over state-of-the-art LFS methods on our large-scale pediatric CXR dataset and also achieving extremely competitive results on adults’ CXR dataset.

Segmentation of organs from chest X-ray images is an essential task for an accurate and reliable diagnosis of lung diseases and chest organ morphometry. In this study, we investigated the benefits of augmenting state-of-the-art deep convolutional neural networks (CNNs) for image segmentation with organ contour information and evaluated the performance of such augmentation on segmentation of lung fields, heart, and clavicles from chest X-ray images. Three state-of-the-art CNNs were augmented, namely the UNet and LinkNet architecture with the ResNeXt feature extraction backbone, and the Tiramisu architecture with the DenseNet. All CNN architectures were trained on ground-truth segmentation masks and additionally on the corresponding contours. The contribution of such contour-based augmentation was evaluated against the contour-free architectures, and 20 existing algorithms for lung field segmentation. The proposed contour-aware segmentation improved the segmentation performance, and when compared against existing algorithms on the same publicly available database of 247 chest X-ray images, the UNet architecture with the ResNeXt50 encoder combined with the contour-aware approach resulted in the best overall segmentation performance, achieving a Jaccard overlap coefficient of 0.971, 0.933, and 0.903 for the lung fields, heart, and clavicles, respectively. In this study, we proposed to augment CNN architectures for CXR segmentation with organ contour information and were able to significantly improve segmentation accuracy and outperform all existing solution using a public chest X-ray database.

Background: Tuberculosis (TB) remains a major global health challenge, causing significant morbidity and mortality. Early and accurate detection is crucial for timely treatment and disease control. Traditional TB diagnosis relies on radiologists analyzing chest X-rays (CXRs), a process that is time-consuming and prone to variability. Advances in artificial intelligence, particularly deep learning, have facilitated the development of computer-aided diagnostic (CAD) systems capable of automating TB detection with improved efficiency and consistency. Objective: This study aimed to develop an automated TB detection system utilizing deep learning techniques to segment lung regions and classify TB-infected CXRs, enhancing diagnostic accuracy and reducing reliance on manual interpretation. Methods: A fully convolutional network (FCN) based segmentation model was implemented to isolate lung regions from CXRs, followed by post-processing techniques to refine segmentation accuracy. The classification module employed a ResNet architecture to differentiate between normal and TB-positive cases. The model was trained and validated on three datasets: the Japanese Society of Radiological Technology (JSRT), Montgomery County (MC), and a locally curated dataset. For classification, the Shenzhen dataset was used. Model performance was evaluated using accuracy, sensitivity, specificity, Dice Similarity Coefficient (DSC), and the area under the curve (AUC). Results: Segmentation accuracy was 97.1% for JSRT, 97.7% for MC, and 94.2% for the local dataset. DSC values were recorded as 95.1%, 95.4%, and 88.0%, respectively. The classification model achieved 84.4% accuracy, with sensitivity of 84.4%, specificity of 90.09%, and AUC of 95.0%. Comparative analysis demonstrated competitive performance with existing methodologies. Conclusion: The proposed deep learning-based CAD system effectively automates TB detection, improving diagnostic efficiency. The integration of advanced segmentation and classification techniques enhances accuracy, facilitating early TB screening. Future research should explore optimizing classification through hybrid deep learning models for improved clinical applicability.

Lung field segmentation for chest radiography is critical to pulmonary disease diagnosis. In this paper, we propose a new deformable model using weighted sparse shape composition with robust initialization to achieve robust and accurate lung field segmentation. Our method consists of three steps: initialization, deformation and regularization. The steps of deformation and regularization are iteratively employed until convergence. First, since a deformable model is sensitive to the initial shape, a robust initialization is obtained by using a novel voting strategy, which allows the reliable patches on the image to vote for each landmark of the initial shape. Then, each point of the initial shape independently deforms to the lung boundary under the guidance of the appearance model, which can distinguish lung tissues from nonlung tissues near the boundary. Finally, the deformed shape is regularized by weighted sparse shape composition (SSC) model, which is constrained by both boundary information and the correlations between each point of the deformed shape. Our method has been evaluated on 247 chest radiographs from well-known dataset Japanese Society of Radiological Technology (JSRT) and achieved high overlap scores (0.955±0.021). The experimental results show that the proposed deformable segmentation model is more robust and accurate than the traditional appearance and shape model on the JSRT database. Our method also shows higher accuracy than most state-of-the-art methods.

Automatic chest anatomy segmentation plays a key role in computer-aided disease diagnosis, such as for cardiomegaly, pleural effusion, emphysema, and pneumothorax. Among these diseases, cardiomegaly is considered a perilous disease, involving a high risk of sudden cardiac death. It can be diagnosed early by an expert medical practitioner using a chest X-Ray (CXR) analysis. The cardiothoracic ratio (CTR) and transverse cardiac diameter (TCD) are the clinical criteria used to estimate the heart size for diagnosing cardiomegaly. Manual estimation of CTR and other diseases is a time-consuming process and requires significant work by the medical expert. Cardiomegaly and related diseases can be automatically estimated by accurate anatomical semantic segmentation of CXRs using artificial intelligence. Automatic segmentation of the lungs and heart from the CXRs is considered an intensive task owing to inferior quality images and intensity variations using nonideal imaging conditions. Although there are a few deep learning-based techniques for chest anatomy segmentation, most of them only consider single class lung segmentation with deep complex architectures that require a lot of trainable parameters. To address these issues, this study presents two multiclass residual mesh-based CXR segmentation networks, X-RayNet-1 and X-RayNet-2, which are specifically designed to provide fine segmentation performance with a few trainable parameters compared to conventional deep learning schemes. The proposed methods utilize semantic segmentation to support the diagnostic procedure of related diseases. To evaluate X-RayNet-1 and X-RayNet-2, experiments were performed with a publicly available Japanese Society of Radiological Technology (JSRT) dataset for multiclass segmentation of the lungs, heart, and clavicle bones; two other publicly available datasets, Montgomery County (MC) and Shenzhen X-Ray sets (SC), were evaluated for lung segmentation. The experimental results showed that X-RayNet-1 achieved fine performance for all datasets and X-RayNet-2 achieved competitive performance with a 75% parameter reduction.

Deep neural networks have entirely dominated the machine vision space in the past few years due to their astonishing human comparable performance. This paper applies power of such network to segment out lungs from chest x-rays, which is a crucial step in any computer aided diagnostic (CAD) system design. A fully convolutional network was used to extract lungs region from the x-rays. Post processing was done to fill holes, separate left and right lung from each other and remove unwanted objects that appeared in few cases. The process was repeated ten times, with random split of data into a 60:40 ratio as training and testing sets respectively, to calculate the average accuracy. The methodology was tested on three datasets: Japanese Society of Radiological Technology (JSRT), Montgomery County (MC), and a local dataset that achieved average accuracy of 97.1%, 97.7% & 94.2% respectively. The results proved that the proposed methodology is efficient enough and can be generalized for other such segmentation problems in medical imaging domain.

A CAD tool for chest radiographs consists of several procedures and the very first step is segmentation of lung fields. We develop a novel methodology for segmentation of lung fields in chest radiographs that can satisfy the following two requirements. First, we aim to develop a segmentation method that does not need a training stage with manual estimation of anatomical features in a large training dataset of images. Secondly, for the ease of implementation, it is desirable to apply a well established model that is widely used for various image-partitioning practices. The Chan-Vese active contour model, which is based on Mumford-Shah functional in the level set framework, is applied for segmentation of lung fields. With the use of this model, segmentation of lung fields can be carried out without detailed prior knowledge on the radiographic anatomy of the chest, yet in some chest radiographs, the trachea regions are unfavorably segmented out in addition to the lung field contours. To eliminate artifacts from the trachea, we locate the upper end of the trachea, find a vertical center line of the trachea and delineate it, and then brighten the trachea region to make it less distinctive. The segmentation process is finalized by subsequent morphological operations. We randomly select 30 images from the Japanese Society of Radiological Technology image database to test the proposed methodology and the results are shown. We hope our segmentation technique can help to promote of CAD tools, especially for emerging chest radiographic imaging techniques such as dual energy radiography and chest tomosynthesis.

A deep unsupervised saliency model for lung segmentation in chest X-ray images

Deep LF-Net: Semantic lung segmentation from Indian chest radiographs including severely unhealthy images

Semantic segmentation for diagnosing chest-related diseases like cardiomegaly, emphysema, pleural effusions, and pneumothorax is a critical yet understudied tool for identifying the chest anatomy. A dangerous disease among these is cardiomegaly, in which sudden death is a high risk. An expert medical practitioner can diagnose cardiomegaly early using a chest radiograph (CXR). Cardiomegaly is a heart enlargement disease that can be analyzed by calculating the transverse cardiac diameter (TCD) and the cardiothoracic ratio (CTR). However, the manual estimation of CTR and other chest-related diseases requires much time from medical experts. Based on their anatomical semantics, artificial intelligence estimates cardiomegaly and related diseases by segmenting CXRs. Unfortunately, due to poor-quality images and variations in intensity, the automatic segmentation of the lungs and heart with CXRs is challenging. Deep learning-based methods are being used to identify the chest anatomy segmentation, but most of them only consider the lung segmentation, requiring a great deal of training. This work is based on a multiclass concatenation-based automatic semantic segmentation network, CardioNet, that was explicitly designed to perform fine segmentation using fewer parameters than a conventional deep learning scheme. Furthermore, the semantic segmentation of other chest-related diseases is diagnosed using CardioNet. CardioNet is evaluated using the JSRT dataset (Japanese Society of Radiological Technology). The JSRT dataset is publicly available and contains multiclass segmentation of the heart, lungs, and clavicle bones. In addition, our study examined lung segmentation using another publicly available dataset, Montgomery County (MC). The experimental results of the proposed CardioNet model achieved acceptable accuracy and competitive results across all datasets.

Segmentation of anatomical structures in Chest Posterior-Anterior Radiographs is a classical task on biomedical image analysis. Deep Learning has been widely used for detection and diagnosis of illnesses in several medical image modalities over the last years, but the portability of deep methods is still limited, hampering the reusability of pre-trained models in new data. We address this problem by proposing a novel method for Cross-Dataset Transfer Learning in Chest X-Ray images based on Unsupervised Image Translation architectures. Our Transfer Learning approach achieved Jaccard values of 88.20% on lung field segmentation in the Montgomery Set by using a pre-trained model on the JSRT dataset and no labeled data from the target dataset. Several experiments in unsupervised and semi-supervised transfer were performed and our method consistently outperformed simple fine-tuning when a limited amount of labels is used. Qualitative analysis on the tasks of clavicle and heart segmentation are also performed on Montgomery samples and pre-trained models from JSRT dataset. Our secondary contributions encompass several experiments in anatomical structure segmentation on JSRT, achieving state-of-the-art results in lung field (96.02%), heart (89.64%) and clavicle segmentation (87.30%).