# OOD Detection on Tiny ImageNet (ID) vs Tiny ObjectNet (OOD) (test)

Source: [OOD Detection on Tiny ImageNet (ID) vs Tiny ObjectNet (OOD) (test)](<https://www.sota2.com/research/sota/ood-detection-on-tiny-imagenet-id-vs-tiny-objectnet-ood-test>)

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

Data updated: 2026-02-26T05:25:49.139Z

Evaluation of Out-of-Distribution detection performance using Tiny ImageNet as the in-distribution dataset and Tiny ObjectNet as the out-of-distribution dataset, measured on the test split.

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 | OOD Detection |
| Dataset (including any stated split) | Tiny ImageNet (ID) vs Tiny ObjectNet (OOD) (test) |
| Metric | AUROC |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [AUROC](<https://www.sota2.com/research/sota/ood-detection-on-tiny-imagenet-id-vs-tiny-objectnet-ood-test.md?metric=AUROC>): Higher is better; unit: SCALAR.

- [TNR @ FPR=95%](<https://www.sota2.com/research/sota/ood-detection-on-tiny-imagenet-id-vs-tiny-objectnet-ood-test.md?metric=TNR%20%40%20FPR%3D95%25>): Higher is better; unit: SCALAR.

## Recorded results

| Method | AUROC | Result date | Report paper |
| --- | --- | --- | --- |
| MCD | 0.99 | Not recorded | 68fa6c001e209131cea3190d |
| ERD++ | 0.99 | Not recorded | 68fa6c001e209131cea3190d |
| ERD | 0.98 | Not recorded | 68fa6c001e209131cea3190d |
| Vanilla Ensembles | 0.82 | Not recorded | 68fa6c001e209131cea3190d |
| OE | 0.79 | Not recorded | 68fa6c001e209131cea3190d |
| Mahal. | 0.75 | Not recorded | 68fa6c001e209131cea3190d |
| Mahal-U | 0.72 | Not recorded | 68fa6c001e209131cea3190d |
| DPN | 0.7 | Not recorded | 68fa6c001e209131cea3190d |

### MCD

AUROC: 0.99. Result ID: 698eb7b0ffc3901de0a00e57.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa665c1e209131cea0fd3b.md>)

Recorded evaluation configuration: {"Backbone":"ResNet20"}. This record does not assert independent reproduction.

### ERD++

AUROC: 0.99. Result ID: 698eb7b0ffc3901de0a00e5a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

[Implementation](<https://github.com/ericpts/reto>)

Recorded evaluation configuration: {"Backbone":"ResNet20","Ensemble size":5}. This record does not assert independent reproduction.

### ERD

AUROC: 0.98. Result ID: 698eb7b0ffc3901de0a00e59.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

[Implementation](<https://github.com/ericpts/reto>)

Recorded evaluation configuration: {"Backbone":"ResNet20","Ensemble size":5}. This record does not assert independent reproduction.

### Vanilla Ensembles

AUROC: 0.82. Result ID: 698eb7b0ffc3901de0a00e53.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63bc1e209131cea0084f.md>)

Recorded evaluation configuration: {"Backbone":"ResNet20","Ensemble size":5}. This record does not assert independent reproduction.

### OE

AUROC: 0.79. Result ID: 698eb7b0ffc3901de0a00e55.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa654c1e209131cea096af.md>)

[Implementation](<https://github.com/hendrycks/outlier-exposure>)

Recorded evaluation configuration: {"Backbone":"ResNet20"}. This record does not assert independent reproduction.

### Mahal.

AUROC: 0.75. Result ID: 698eb7b0ffc3901de0a00e56.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64f91e209131cea077ac.md>)

[Implementation](<https://github.com/pokaxpoka/deep_Mahalanobis_detector>)

Recorded evaluation configuration: {"Backbone":"ResNet20"}. This record does not assert independent reproduction.

### Mahal-U

AUROC: 0.72. Result ID: 698eb7b0ffc3901de0a00e58.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

Recorded evaluation configuration: {"Backbone":"ResNet20"}. This record does not assert independent reproduction.

### DPN

AUROC: 0.7. Result ID: 698eb7b0ffc3901de0a00e54.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c001e209131cea3190d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa651f1e209131cea088e1.md>)

Recorded evaluation configuration: {"Backbone":"ResNet20"}. This record does not assert independent reproduction.

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