# Image Classification on EuroSAT (Adversarial Robustness Metrics)

Source: [Image Classification on EuroSAT (Adversarial Robustness Metrics)](<https://www.sota2.com/research/sota/image-classification-on-eurosat-adversarial-robustness-metrics>)

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

Data updated: 2026-04-23T07:35:04.220Z

Evaluation of Image Classification performance on the EuroSAT dataset focusing on adversarial robustness metrics, including Clean Accuracy, PGD, and AutoAttack results.

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 | Image Classification |
| Dataset (including any stated split) | EuroSAT |
| Metric | Robust Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 21 of 21 |
| 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

- [Robust Accuracy](<https://www.sota2.com/research/sota/image-classification-on-eurosat-adversarial-robustness-metrics.md?metric=Robust%20Accuracy>): Higher is better; unit: PERCENT.

- [Clean Accuracy](<https://www.sota2.com/research/sota/image-classification-on-eurosat-adversarial-robustness-metrics.md?metric=Clean%20Accuracy>): Higher is better; unit: PERCENT.

- [AutoAttack Error](<https://www.sota2.com/research/sota/image-classification-on-eurosat-adversarial-robustness-metrics.md?metric=AutoAttack%20Error>): Lower is better; unit: SCALAR.

- [PGD Error Rate](<https://www.sota2.com/research/sota/image-classification-on-eurosat-adversarial-robustness-metrics.md?metric=PGD%20Error%20Rate>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Robust Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| CSR | 26.1 | Not recorded | 69799806a2b28e761d87cb52 |
| R-TPT | 21.4 | Not recorded | 69799806a2b28e761d87cb52 |
| Ours (5 trees) | 18.89 | Not recorded | 69e85bd47d0172140ec81123 |
| Ours | 17.49 | Not recorded | 69e85bd47d0172140ec81123 |
| FARE | 17.4 | Not recorded | 69799806a2b28e761d87cb52 |
| AoS | 17.32 | Not recorded | 69e85bd47d0172140ec81123 |
| TeCoA | 16.58 | Not recorded | 69e85bd47d0172140ec81123 |
| PMG-FT | 14.94 | Not recorded | 69e85bd47d0172140ec81123 |
| TeCoA | 14.4 | Not recorded | 69799806a2b28e761d87cb52 |
| TTC | 13.9 | Not recorded | 69799806a2b28e761d87cb52 |
| Anti-Adv | 13.7 | Not recorded | 69799806a2b28e761d87cb52 |
| TTE | 10.2 | Not recorded | 69799806a2b28e761d87cb52 |
| FARE | 10.15 | Not recorded | 69e85bd47d0172140ec81123 |
| HD | 2.1 | Not recorded | 69799806a2b28e761d87cb52 |
| LPF | 1.5 | Not recorded | 69799806a2b28e761d87cb52 |
| CLIP | 0.05 | Not recorded | 69e85bd47d0172140ec81123 |
| CLIP | 0 | Not recorded | 69799806a2b28e761d87cb52 |

### CSR

Robust Accuracy: 26.1. Result ID: 69a17d1735dfcabb00e7619a.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Test-Time Defense"}. This record does not assert independent reproduction.

### R-TPT

Robust Accuracy: 21.4. Result ID: 69a17d1735dfcabb00e76194.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7add1e209131cea8590d.md>)

[Implementation](<https://github.com/TomSheng21/R-TPT>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Test-Time Defense"}. This record does not assert independent reproduction.

### Ours (5 trees)

Robust Accuracy: 18.89. Result ID: 69e9cb8a47f34ce924d09e56.

[Reporting paper](<https://www.sota2.com/research/paper/69e85bd47d0172140ec81123.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e85bd47d0172140ec81123.md>)

Recorded evaluation configuration: {"Zero-shot":"true","Adversarial fine-tuning":"ImageNet","PGD attack steps":"20","Perturbation radius (epsilon)":"1/255","Number of trees":"5"}. This record does not assert independent reproduction.

### Ours

Robust Accuracy: 17.49. Result ID: 69e9cb8a47f34ce924d09e55.

[Reporting paper](<https://www.sota2.com/research/paper/69e85bd47d0172140ec81123.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e85bd47d0172140ec81123.md>)

Recorded evaluation configuration: {"Zero-shot":"true","Adversarial fine-tuning":"ImageNet","PGD attack steps":"20","Perturbation radius (epsilon)":"1/255"}. This record does not assert independent reproduction.

### FARE

Robust Accuracy: 17.4. Result ID: 69a17d1735dfcabb00e76193.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738c1e209131cea5d24a.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Adversarial Fine-tuning"}. This record does not assert independent reproduction.

### AoS

Robust Accuracy: 17.32. Result ID: 69e9cb8a47f34ce924d09e54.

[Reporting paper](<https://www.sota2.com/research/paper/69e85bd47d0172140ec81123.md>)

Recorded evaluation configuration: {"Zero-shot":"true","Adversarial fine-tuning":"ImageNet","PGD attack steps":"20","Perturbation radius (epsilon)":"1/255"}. This record does not assert independent reproduction.

### TeCoA

Robust Accuracy: 16.58. Result ID: 69e9cb8a47f34ce924d09e51.

[Reporting paper](<https://www.sota2.com/research/paper/69e85bd47d0172140ec81123.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f271e209131cea44177.md>)

Recorded evaluation configuration: {"Zero-shot":"true","Adversarial fine-tuning":"ImageNet","PGD attack steps":"20","Perturbation radius (epsilon)":"1/255"}. This record does not assert independent reproduction.

### PMG-FT

Robust Accuracy: 14.94. Result ID: 69e9cb8a47f34ce924d09e52.

[Reporting paper](<https://www.sota2.com/research/paper/69e85bd47d0172140ec81123.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72df1e209131cea5912e.md>)

Recorded evaluation configuration: {"Zero-shot":"true","Adversarial fine-tuning":"ImageNet","PGD attack steps":"20","Perturbation radius (epsilon)":"1/255"}. This record does not assert independent reproduction.

### TeCoA

Robust Accuracy: 14.4. Result ID: 69a17d1735dfcabb00e76192.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f271e209131cea44177.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Adversarial Fine-tuning"}. This record does not assert independent reproduction.

### TTC

Robust Accuracy: 13.9. Result ID: 69a17d1735dfcabb00e76199.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa774a1e209131cea724cc.md>)

[Implementation](<https://www.sota2.com/here>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Test-Time Defense"}. This record does not assert independent reproduction.

### Anti-Adv

Robust Accuracy: 13.7. Result ID: 69a17d1735dfcabb00e76197.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b7f1e209131cea2e9af.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Test-Time Defense"}. This record does not assert independent reproduction.

### TTE

Robust Accuracy: 10.2. Result ID: 69a17d1735dfcabb00e76198.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a711e209131cea2817e.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Test-Time Defense"}. This record does not assert independent reproduction.

### FARE

Robust Accuracy: 10.15. Result ID: 69e9cb8a47f34ce924d09e53.

[Reporting paper](<https://www.sota2.com/research/paper/69e85bd47d0172140ec81123.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738c1e209131cea5d24a.md>)

Recorded evaluation configuration: {"Zero-shot":"true","Adversarial fine-tuning":"ImageNet","PGD attack steps":"20","Perturbation radius (epsilon)":"1/255"}. This record does not assert independent reproduction.

### HD

Robust Accuracy: 2.1. Result ID: 69a17d1735dfcabb00e76196.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a1d1e209131cea26045.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Test-Time Defense"}. This record does not assert independent reproduction.

### LPF

Robust Accuracy: 1.5. Result ID: 69a17d1735dfcabb00e76195.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Test-Time Defense"}. This record does not assert independent reproduction.

### CLIP

Robust Accuracy: 0.05. Result ID: 69e9cb8a47f34ce924d09e50.

[Reporting paper](<https://www.sota2.com/research/paper/69e85bd47d0172140ec81123.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Zero-shot":"true","Adversarial fine-tuning":"ImageNet","PGD attack steps":"20","Perturbation radius (epsilon)":"1/255"}. This record does not assert independent reproduction.

### CLIP

Robust Accuracy: 0. Result ID: 69a17d1735dfcabb00e76191.

[Reporting paper](<https://www.sota2.com/research/paper/69799806a2b28e761d87cb52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Backbone":"CLIP-B/32","Category":"Original"}. This record does not assert independent reproduction.

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