# Image Classification Calibration on ImageNet Sketch

Source: [Image Classification Calibration on ImageNet Sketch](<https://www.sota2.com/research/sota/image-classification-calibration-on-imagenet-sketch>)

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

Data updated: 2026-05-22T07:15:58.605Z

Evaluation of model calibration performance on the ImageNet-Sketch dataset, measured using metrics like Expected Calibration Error (ECE) and accuracy difference (ΔACC).

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 Calibration |
| Dataset (including any stated split) | ImageNet Sketch |
| Metric | CwECE |
| Unit | SCALAR |
| Direction | Lower 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

- [CwECE](<https://www.sota2.com/research/sota/image-classification-calibration-on-imagenet-sketch.md?metric=CwECE>): Lower is better; unit: SCALAR.

- [ECE](<https://www.sota2.com/research/sota/image-classification-calibration-on-imagenet-sketch.md?metric=ECE>): Lower is better; unit: SCALAR.

- [Accuracy Difference (ΔACC)](<https://www.sota2.com/research/sota/image-classification-calibration-on-imagenet-sketch.md?metric=Accuracy%20Difference%20(%CE%94ACC)>): Lower is better; unit: SCALAR.

- [ECEKDE](<https://www.sota2.com/research/sota/image-classification-calibration-on-imagenet-sketch.md?metric=ECEKDE>): Lower is better; unit: SCALAR.

## Recorded results

| Method | CwECE | Result date | Report paper |
| --- | --- | --- | --- |
| Oracle | 0.93 | Not recorded | 6a0fe2e3f984739ed6739baf |
| Oracle | 1.64 | Not recorded | 6a0fe2e3f984739ed6739baf |
| Oracle | 1.76 | Not recorded | 6a0fe2e3f984739ed6739baf |
| ECL | 5.18 | Not recorded | 6a0fe2e3f984739ed6739baf |
| ECL | 9.86 | Not recorded | 6a0fe2e3f984739ed6739baf |
| ECL | 10.7 | Not recorded | 6a0fe2e3f984739ed6739baf |

### Oracle

CwECE: 0.93. Result ID: 6a10025ffbc373592c087d0f.

[Reporting paper](<https://www.sota2.com/research/paper/6a0fe2e3f984739ed6739baf.md>)

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

### Oracle

CwECE: 1.64. Result ID: 6a10025ffbc373592c087d0b.

[Reporting paper](<https://www.sota2.com/research/paper/6a0fe2e3f984739ed6739baf.md>)

Recorded evaluation configuration: {"Backbone":"ResNet152"}. This record does not assert independent reproduction.

### Oracle

CwECE: 1.76. Result ID: 6a10025ffbc373592c087d0d.

[Reporting paper](<https://www.sota2.com/research/paper/6a0fe2e3f984739ed6739baf.md>)

Recorded evaluation configuration: {"Backbone":"DenseNet161"}. This record does not assert independent reproduction.

### ECL

CwECE: 5.18. Result ID: 6a10025ffbc373592c087d0e.

[Reporting paper](<https://www.sota2.com/research/paper/6a0fe2e3f984739ed6739baf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0fe2e3f984739ed6739baf.md>)

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

### ECL

CwECE: 9.86. Result ID: 6a10025ffbc373592c087d0a.

[Reporting paper](<https://www.sota2.com/research/paper/6a0fe2e3f984739ed6739baf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0fe2e3f984739ed6739baf.md>)

Recorded evaluation configuration: {"Backbone":"ResNet152"}. This record does not assert independent reproduction.

### ECL

CwECE: 10.7. Result ID: 6a10025ffbc373592c087d0c.

[Reporting paper](<https://www.sota2.com/research/paper/6a0fe2e3f984739ed6739baf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0fe2e3f984739ed6739baf.md>)

Recorded evaluation configuration: {"Backbone":"DenseNet161"}. This record does not assert independent reproduction.

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