# Thoracic Pathology Classification on ChestX-ray14 2017 (test)

Source: [Thoracic Pathology Classification on ChestX-ray14 2017 (test)](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test>)

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

Data updated: 2026-02-26T05:51:46.425Z

Evaluation of multi-label classification performance for 14 specific thoracic pathologies on the test split of the ChestX-ray14 2017 dataset.

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 | Thoracic Pathology Classification |
| Dataset (including any stated split) | ChestX-ray14 2017 (test) |
| Metric | Mean AUC |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 32 of 32 |
| 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

- [Mean AUC](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Mean%20AUC>): Higher is better; unit: PERCENT.

- [Atelectasis](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Atelectasis>): Higher is better; unit: PERCENT.

- [Cardiomegaly](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Cardiomegaly>): Higher is better; unit: PERCENT.

- [Effusion](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Effusion>): Higher is better; unit: PERCENT.

- [Infiltration](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Infiltration>): Higher is better; unit: PERCENT.

- [Mass](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Mass>): Higher is better; unit: PERCENT.

- [Nodule](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Nodule>): Higher is better; unit: PERCENT.

- [Pneumonia](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Pneumonia>): Higher is better; unit: PERCENT.

- [Pneumothorax](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Pneumothorax>): Higher is better; unit: PERCENT.

- [Consolidation](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Consolidation>): Higher is better; unit: PERCENT.

- [Edema](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Edema>): Higher is better; unit: PERCENT.

- [Emphysema](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Emphysema>): Higher is better; unit: PERCENT.

- [Fibrosis](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Fibrosis>): Higher is better; unit: PERCENT.

- [Pleural Thickening](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Pleural%20Thickening>): Higher is better; unit: PERCENT.

- [Hernia](<https://www.sota2.com/research/sota/thoracic-pathology-classification-on-chestx-ray14-2017-test.md?metric=Hernia>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Mean AUC | Result date | Report paper |
| --- | --- | --- | --- |
| ACPL | 81.77 | Not recorded | 68fa6c0c1e209131cea31acf |
| S2MTS2 | 81.06 | Not recorded | 68fa6c0c1e209131cea31acf |
| ACPL | 81.06 | Not recorded | 68fa6c0c1e209131cea31acf |
| ACPL | 80.4 | Not recorded | 68fa6c0c1e209131cea31acf |
| S2MTS2 | 80.31 | Not recorded | 68fa6c0c1e209131cea31acf |
| UPS | 79.92 | Not recorded | 68fa6c0c1e209131cea31acf |
| S2MTS2 | 79.9 | Not recorded | 68fa6c0c1e209131cea31acf |
| NoTeacher | 79.49 | Not recorded | 68fa6c0c1e209131cea31acf |
| SRC-MT | 79.23 | Not recorded | 68fa6c0c1e209131cea31acf |
| ACPL | 79.2 | Not recorded | 68fa6c0c1e209131cea31acf |
| S2MTS2 | 78.96 | Not recorded | 68fa6c0c1e209131cea31acf |
| UPS | 78.9 | Not recorded | 68fa6c0c1e209131cea31acf |
| Graph XNet | 78 | Not recorded | 68fa6c0c1e209131cea31acf |
| SRC-MT | 77.76 | Not recorded | 68fa6c0c1e209131cea31acf |
| NoTeacher | 77.61 | Not recorded | 68fa6c0c1e209131cea31acf |
| NoTeacher | 77.04 | Not recorded | 68fa6c0c1e209131cea31acf |
| UPS | 76.84 | Not recorded | 68fa6c0c1e209131cea31acf |
| SRC-MT | 75.28 | Not recorded | 68fa6c0c1e209131cea31acf |
| ACPL | 74.82 | Not recorded | 68fa6c0c1e209131cea31acf |
| S2MTS2 | 74.69 | Not recorded | 68fa6c0c1e209131cea31acf |
| UPS | 73.18 | Not recorded | 68fa6c0c1e209131cea31acf |
| NoTeacher | 72.6 | Not recorded | 68fa6c0c1e209131cea31acf |
| SRC-MT | 72.29 | Not recorded | 68fa6c0c1e209131cea31acf |
| Graph XNet | 68 | Not recorded | 68fa6c0c1e209131cea31acf |
| SRC-MT | 66.95 | Not recorded | 68fa6c0c1e209131cea31acf |
| UPS | 65.51 | Not recorded | 68fa6c0c1e209131cea31acf |
| Graph XNet | 63 | Not recorded | 68fa6c0c1e209131cea31acf |
| Graph XNet | 58 | Not recorded | 68fa6c0c1e209131cea31acf |
| Graph XNet | 53 | Not recorded | 68fa6c0c1e209131cea31acf |

### ACPL

Mean AUC: 81.77. Result ID: 698ebc16ffc3901de0a034d3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Implementation](<https://github.com/FBLADL/ACPL>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"20%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### S2MTS2

Mean AUC: 81.06. Result ID: 698ebc16ffc3901de0a034c4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2a1e209131cea2c73f.md>)

[Implementation](<https://github.com/fengbeiliu/semi-chest>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"20%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### ACPL

Mean AUC: 81.06. Result ID: 698ebc16ffc3901de0a034d2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Implementation](<https://github.com/FBLADL/ACPL>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"15%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### ACPL

Mean AUC: 80.4. Result ID: 698ebc16ffc3901de0a034d1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Implementation](<https://github.com/FBLADL/ACPL>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"10%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### S2MTS2

Mean AUC: 80.31. Result ID: 698ebc16ffc3901de0a034c3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2a1e209131cea2c73f.md>)

[Implementation](<https://github.com/fengbeiliu/semi-chest>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"15%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### UPS

Mean AUC: 79.92. Result ID: 698ebc16ffc3901de0a034ce.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69ce1e209131cea2435a.md>)

[Implementation](<https://github.com/nayeemrizve/ups>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"20%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### S2MTS2

Mean AUC: 79.9. Result ID: 698ebc16ffc3901de0a034c2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2a1e209131cea2c73f.md>)

[Implementation](<https://github.com/fengbeiliu/semi-chest>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"10%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### NoTeacher

Mean AUC: 79.49. Result ID: 698ebc16ffc3901de0a034bf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"20%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### SRC-MT

Mean AUC: 79.23. Result ID: 698ebc16ffc3901de0a034bb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa678c1e209131cea16e1f.md>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"20%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

### ACPL

Mean AUC: 79.2. Result ID: 698ebc16ffc3901de0a034d0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Implementation](<https://github.com/FBLADL/ACPL>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"5%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### S2MTS2

Mean AUC: 78.96. Result ID: 698ebc16ffc3901de0a034c1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2a1e209131cea2c73f.md>)

[Implementation](<https://github.com/fengbeiliu/semi-chest>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"5%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### UPS

Mean AUC: 78.9. Result ID: 698ebc16ffc3901de0a034cd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69ce1e209131cea2435a.md>)

[Implementation](<https://github.com/nayeemrizve/ups>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"15%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### Graph XNet

Mean AUC: 78. Result ID: 698ebc16ffc3901de0a034c9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67ed1e209131cea19350.md>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"20%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

### SRC-MT

Mean AUC: 77.76. Result ID: 698ebc16ffc3901de0a034ba.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa678c1e209131cea16e1f.md>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"15%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

### NoTeacher

Mean AUC: 77.61. Result ID: 698ebc16ffc3901de0a034be.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"10%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### NoTeacher

Mean AUC: 77.04. Result ID: 698ebc16ffc3901de0a034bd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"5%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### UPS

Mean AUC: 76.84. Result ID: 698ebc16ffc3901de0a034cc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69ce1e209131cea2435a.md>)

[Implementation](<https://github.com/nayeemrizve/ups>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"10%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### SRC-MT

Mean AUC: 75.28. Result ID: 698ebc16ffc3901de0a034b9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa678c1e209131cea16e1f.md>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"10%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

### ACPL

Mean AUC: 74.82. Result ID: 698ebc16ffc3901de0a034cf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Implementation](<https://github.com/FBLADL/ACPL>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"2%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### S2MTS2

Mean AUC: 74.69. Result ID: 698ebc16ffc3901de0a034c0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2a1e209131cea2c73f.md>)

[Implementation](<https://github.com/fengbeiliu/semi-chest>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"2%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### UPS

Mean AUC: 73.18. Result ID: 698ebc16ffc3901de0a034cb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69ce1e209131cea2435a.md>)

[Implementation](<https://github.com/nayeemrizve/ups>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"5%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### NoTeacher

Mean AUC: 72.6. Result ID: 698ebc16ffc3901de0a034bc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"2%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### SRC-MT

Mean AUC: 72.29. Result ID: 698ebc16ffc3901de0a034b8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa678c1e209131cea16e1f.md>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"5%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

### Graph XNet

Mean AUC: 68. Result ID: 698ebc16ffc3901de0a034c8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67ed1e209131cea19350.md>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"15%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

### SRC-MT

Mean AUC: 66.95. Result ID: 698ebc16ffc3901de0a034b7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa678c1e209131cea16e1f.md>)

Recorded evaluation configuration: {"Method Type":"Consistency based","Label Percentage":"2%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

### UPS

Mean AUC: 65.51. Result ID: 698ebc16ffc3901de0a034ca.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69ce1e209131cea2435a.md>)

[Implementation](<https://github.com/nayeemrizve/ups>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"2%","Backbone":"DenseNet-121"}. This record does not assert independent reproduction.

### Graph XNet

Mean AUC: 63. Result ID: 698ebc16ffc3901de0a034c7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67ed1e209131cea19350.md>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"10%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

### Graph XNet

Mean AUC: 58. Result ID: 698ebc16ffc3901de0a034c6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67ed1e209131cea19350.md>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"5%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

### Graph XNet

Mean AUC: 53. Result ID: 698ebc16ffc3901de0a034c5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c0c1e209131cea31acf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67ed1e209131cea19350.md>)

Recorded evaluation configuration: {"Method Type":"Pseudo Label","Label Percentage":"2%","Backbone":"DenseNet-169"}. This record does not assert independent reproduction.

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