# Lung Segmentation on PadChest (five-fold cross-validation)

Source: [Lung Segmentation on PadChest (five-fold cross-validation)](<https://www.sota2.com/research/sota/lung-segmentation-on-padchest-five-fold-cross-validation>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-19T08:05:00.452Z

Evaluation of Lung Segmentation performance on the PadChest dataset using a five-fold cross-validation strategy, measured using Dice Similarity Coefficients (DSC), Hausdorff Distance (HD95), and Average Surface Distance (ASD).

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Lung segmentation |
| Dataset (including any stated split) | PadChest (five-fold cross-validation) |
| Metric | DSC (%) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [DSC (%)](<https://www.sota2.com/research/sota/lung-segmentation-on-padchest-five-fold-cross-validation.md?metric=DSC%20(%25)>): Higher is better; unit: PERCENT.

- [DSC (NF) (%)](<https://www.sota2.com/research/sota/lung-segmentation-on-padchest-five-fold-cross-validation.md?metric=DSC%20(NF)%20(%25)>): Higher is better; unit: PERCENT.

- [DSC (PE) (%)](<https://www.sota2.com/research/sota/lung-segmentation-on-padchest-five-fold-cross-validation.md?metric=DSC%20(PE)%20(%25)>): Higher is better; unit: PERCENT.

- [HD95](<https://www.sota2.com/research/sota/lung-segmentation-on-padchest-five-fold-cross-validation.md?metric=HD95>): Lower is better; unit: SCALAR.

- [ASD](<https://www.sota2.com/research/sota/lung-segmentation-on-padchest-five-fold-cross-validation.md?metric=ASD>): Lower is better; unit: SCALAR.

## Recorded results

| Method | DSC (%) | Result date | Report paper |
| --- | --- | --- | --- |
| S-MVD-CL | 93.93 | Not recorded | 69bb82ee0636e5bf4afc18c6 |
| CheXMask | 93.65 | Not recorded | 69bb82ee0636e5bf4afc18c6 |
| S-DVD-CL | 93.15 | Not recorded | 69bb82ee0636e5bf4afc18c6 |
| SSDCL | 92.91 | Not recorded | 69bb82ee0636e5bf4afc18c6 |
| DVD-CL | 92.76 | Not recorded | 69bb82ee0636e5bf4afc18c6 |
| MVD-CL | 92.27 | Not recorded | 69bb82ee0636e5bf4afc18c6 |
| VADeR | 92.26 | Not recorded | 69bb82ee0636e5bf4afc18c6 |
| SimCLR | 90.9 | Not recorded | 69bb82ee0636e5bf4afc18c6 |
| None | 87.05 | Not recorded | 69bb82ee0636e5bf4afc18c6 |

### S-MVD-CL

DSC (%): 93.93. Result ID: 69bbae0e47f34ce924a55eb0.

[Reporting paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

Recorded evaluation configuration: {"Pre-training Method":"S-MVD-CL","Supervision type":"Supervised","Fine-tuning evaluation protocol":"five-fold cross-validation"}. This record does not assert independent reproduction.

### CheXMask

DSC (%): 93.65. Result ID: 69bbae0e47f34ce924a55ead.

[Reporting paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa471a1e209131ce92576e.md>)

Recorded evaluation configuration: {"Pre-training Method":"CheXMask","Supervision type":"Supervised","Fine-tuning evaluation protocol":"five-fold cross-validation"}. This record does not assert independent reproduction.

### S-DVD-CL

DSC (%): 93.15. Result ID: 69bbae0e47f34ce924a55eaf.

[Reporting paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

Recorded evaluation configuration: {"Pre-training Method":"S-DVD-CL","Supervision type":"Supervised","Fine-tuning evaluation protocol":"five-fold cross-validation"}. This record does not assert independent reproduction.

### SSDCL

DSC (%): 92.91. Result ID: 69bbae0e47f34ce924a55eae.

[Reporting paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6aea1e209131cea2b006.md>)

[Implementation](<https://github.com/xhu248/semi_cotrast_seg>)

Recorded evaluation configuration: {"Pre-training Method":"SSDCL","Supervision type":"Supervised","Fine-tuning evaluation protocol":"five-fold cross-validation"}. This record does not assert independent reproduction.

### DVD-CL

DSC (%): 92.76. Result ID: 69bbae0e47f34ce924a55eab.

[Reporting paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

Recorded evaluation configuration: {"Pre-training Method":"DVD-CL","Supervision type":"Unsupervised","Fine-tuning evaluation protocol":"five-fold cross-validation"}. This record does not assert independent reproduction.

### MVD-CL

DSC (%): 92.27. Result ID: 69bbae0e47f34ce924a55eac.

[Reporting paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

Recorded evaluation configuration: {"Pre-training Method":"MVD-CL","Supervision type":"Unsupervised","Fine-tuning evaluation protocol":"five-fold cross-validation"}. This record does not assert independent reproduction.

### VADeR

DSC (%): 92.26. Result ID: 69bbae0e47f34ce924a55eaa.

[Reporting paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68e21e209131cea1ef6d.md>)

Recorded evaluation configuration: {"Pre-training Method":"VADeR","Supervision type":"Unsupervised","Fine-tuning evaluation protocol":"five-fold cross-validation"}. This record does not assert independent reproduction.

### SimCLR

DSC (%): 90.9. Result ID: 69bbae0e47f34ce924a55ea9.

[Reporting paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67d71e209131cea18de5.md>)

[Implementation](<https://github.com/google-research/simclr>)

Recorded evaluation configuration: {"Pre-training Method":"SimCLR","Supervision type":"Unsupervised","Fine-tuning evaluation protocol":"five-fold cross-validation"}. This record does not assert independent reproduction.

### None

DSC (%): 87.05. Result ID: 69bbae0e47f34ce924a55ea8.

[Reporting paper](<https://www.sota2.com/research/paper/69bb82ee0636e5bf4afc18c6.md>)

Recorded evaluation configuration: {"Pre-training Method":"None","Supervision type":"Baseline","Fine-tuning evaluation protocol":"five-fold cross-validation"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
