Papers › SPCXR: Self-supervised Pretraining using Chest X-rays Towards a Domain Specific...

SPCXR: Self-supervised Pretraining using Chest X-rays Towards a Domain Specific Foundation Model

23 Nov 2022arXiv:2211.12944archive 2025-07-28

Syed Muhammad Anwar, Abhijeet Parida, Sara Atito, Muhammad Awais, Gustavo Nino, Josef Kitler, Marius George Linguraru

Chest X-rays (CXRs) are a widely used imaging modality for the diagnosis and prognosis of lung disease. The image analysis tasks vary. Examples include pathology detection and lung segmentation. There is a large body of work where machine learning algorithms are developed for specific tasks. A significant recent example is Coronavirus disease (covid-19) detection using CXR data. However, the traditional diagnostic tool design methods based on supervised learning are burdened by the need to provide training data annotation, which should be of good quality for better clinical outcomes. Here, we propose an alternative solution, a new self-supervised paradigm, where a general representation from CXRs is learned using a group-masked self-supervised framework. The pre-trained model is then fine-tuned for domain-specific tasks such as covid-19, pneumonia detection, and general health screening. We show that the same pre-training can be used for the lung segmentation task. Our proposed paradigm shows robust performance in multiple downstream tasks which demonstrates the success of the pre-training. Moreover, the performance of the pre-trained models on data with significant drift during test time proves the learning of a better generic representation. The methods are further validated by covid-19 detection in a unique small-scale pediatric data set. The performance gain in accuracy (~25%) is significant when compared to a supervised transformer-based method. This adds credence to the strength and reliability of our proposed framework and pre-training strategy.

PaperPDF

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

COVID-19 DiagnosisDiagnosticImage SegmentationPneumonia DetectionPrognosisRepresentation LearningSemantic Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
COVID-19 Diagnosis Thoracic Radiologist Per-Class Accuracy 64.37 #2 of 2 Archive leaderboard report
COVID-19 Diagnosis COVIDx CXR-3 SS-CXR Per-Class Accuracy 98.25 #1 of 7 Archive leaderboard report
COVID-19 Diagnosis COVIDx CXR-3 DenseNet-169 Per-Class Accuracy 98.15 #2 of 7 Archive leaderboard report
COVID-19 Diagnosis COVIDx CXR-3 EfficientNet-B2 Per-Class Accuracy 97.6 #3 of 7 Archive leaderboard report
COVID-19 Diagnosis COVIDx CXR-3 Inception Resnet V2 Per-Class Accuracy 97.55 #4 of 7 Archive leaderboard report
COVID-19 Diagnosis COVIDx CXR-3 Inception ResNet Per-Class Accuracy 97.5 #5 of 7 Archive leaderboard report
COVID-19 Diagnosis COVIDx CXR-3 DenseNet-121 Per-Class Accuracy 96.5 #6 of 7 Archive leaderboard report
COVID-19 Diagnosis COVIDx CXR-3 ViT-S Per-Class Accuracy 89.25 #7 of 7 Archive leaderboard report
Semantic Segmentation Montgomery County X-ray Set UNETR + SS-CXR F1-score 0.9561 #1 of 3 Archive leaderboard report
Semantic Segmentation Montgomery County X-ray Set UNETR+ SS-IN F1-score 0.9453 #2 of 3 Archive leaderboard report
Semantic Segmentation Montgomery County X-ray Set UNETR F1-score 0.9227 #3 of 3 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

1x1 ConvolutionAttentionAttention DropoutBatch NormalizationConcatenated Skip ConnectionConvolutionDeiTDense ConnectionsDropoutFeedforward NetworkLinear LayerMax PoolingMulti-Head AttentionPosition-Wise Feed-Forward LayerReLUResidual ConnectionSoftmaxTestU-NetUNETR

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections