# Semantic Segmentation on NWPU Traffic (test)

Source: [Semantic Segmentation on NWPU Traffic (test)](<https://www.sota2.com/research/sota/semantic-segmentation-on-nwpu-traffic-test>)

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

Data updated: 2026-07-07T09:51:21.716Z

Evaluation of Semantic Segmentation performance on the test split of the NWPU Traffic dataset, reporting metrics including mIoU, OA, and F1-score.

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 | Semantic Segmentation |
| Dataset (including any stated split) | NWPU Traffic (test) |
| Metric | Background Accuracy |
| Unit | PERCENT |
| 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

- [Background Accuracy](<https://www.sota2.com/research/sota/semantic-segmentation-on-nwpu-traffic-test.md?metric=Background%20Accuracy>): Higher is better; unit: PERCENT.

- [Airplane Accuracy](<https://www.sota2.com/research/sota/semantic-segmentation-on-nwpu-traffic-test.md?metric=Airplane%20Accuracy>): Higher is better; unit: PERCENT.

- [Car Accuracy](<https://www.sota2.com/research/sota/semantic-segmentation-on-nwpu-traffic-test.md?metric=Car%20Accuracy>): Higher is better; unit: PERCENT.

- [Ship Accuracy](<https://www.sota2.com/research/sota/semantic-segmentation-on-nwpu-traffic-test.md?metric=Ship%20Accuracy>): Higher is better; unit: PERCENT.

- [Train Accuracy](<https://www.sota2.com/research/sota/semantic-segmentation-on-nwpu-traffic-test.md?metric=Train%20Accuracy>): Higher is better; unit: PERCENT.

- [mIoU](<https://www.sota2.com/research/sota/semantic-segmentation-on-nwpu-traffic-test.md?metric=mIoU>): Higher is better; unit: PERCENT.

- [OA](<https://www.sota2.com/research/sota/semantic-segmentation-on-nwpu-traffic-test.md?metric=OA>): Higher is better; unit: PERCENT.

- [F1 Score](<https://www.sota2.com/research/sota/semantic-segmentation-on-nwpu-traffic-test.md?metric=F1%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Background Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| CSPNet | 98.73 | Not recorded | 6a4c882885dba69fef8b29de |
| Deeplabv3+ | 98.66 | Not recorded | 6a4c882885dba69fef8b29de |
| CMLFormer | 98.64 | Not recorded | 6a4c882885dba69fef8b29de |
| GCBNet | 98.48 | Not recorded | 6a4c882885dba69fef8b29de |
| FSEL | 98.47 | Not recorded | 6a4c882885dba69fef8b29de |
| Scaleformer | 98.22 | Not recorded | 6a4c882885dba69fef8b29de |
| UNetFormer | 97.98 | Not recorded | 6a4c882885dba69fef8b29de |
| ABCNet | 97.86 | Not recorded | 6a4c882885dba69fef8b29de |

### CSPNet

Background Accuracy: 98.73. Result ID: 6a4ccc0260b88bb339c9f0a4.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c882885dba69fef8b29de.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a4c882885dba69fef8b29de.md>)

Recorded evaluation configuration: {"FLOPs (G)":"118.7","Params (M)":"29.6"}. This record does not assert independent reproduction.

### Deeplabv3+

Background Accuracy: 98.66. Result ID: 6a4ccc0260b88bb339c9f09d.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c882885dba69fef8b29de.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64b91e209131cea06121.md>)

[Implementation](<https://github.com/tensorflow/models/tree/master/research/deeplab>)

Recorded evaluation configuration: {"FLOPs (G)":"185.6","Params (M)":"41.2"}. This record does not assert independent reproduction.

### CMLFormer

Background Accuracy: 98.64. Result ID: 6a4ccc0260b88bb339c9f0a2.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c882885dba69fef8b29de.md>)

Recorded evaluation configuration: {"FLOPs (G)":"148.5","Params (M)":"38.4"}. This record does not assert independent reproduction.

### GCBNet

Background Accuracy: 98.48. Result ID: 6a4ccc0260b88bb339c9f0a0.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c882885dba69fef8b29de.md>)

Recorded evaluation configuration: {"FLOPs (G)":"156.8","Params (M)":"35.7"}. This record does not assert independent reproduction.

### FSEL

Background Accuracy: 98.47. Result ID: 6a4ccc0260b88bb339c9f0a3.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c882885dba69fef8b29de.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74ab1e209131cea630d9.md>)

Recorded evaluation configuration: {"FLOPs (G)":"126.9","Params (M)":"31.5"}. This record does not assert independent reproduction.

### Scaleformer

Background Accuracy: 98.22. Result ID: 6a4ccc0260b88bb339c9f0a1.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c882885dba69fef8b29de.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d221e209131cea380a3.md>)

Recorded evaluation configuration: {"FLOPs (G)":"134.2","Params (M)":"32.8"}. This record does not assert independent reproduction.

### UNetFormer

Background Accuracy: 97.98. Result ID: 6a4ccc0260b88bb339c9f09f.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c882885dba69fef8b29de.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea362b4.md>)

Recorded evaluation configuration: {"FLOPs (G)":"112.7","Params (M)":"30.1"}. This record does not assert independent reproduction.

### ABCNet

Background Accuracy: 97.86. Result ID: 6a4ccc0260b88bb339c9f09e.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c882885dba69fef8b29de.md>)

Recorded evaluation configuration: {"FLOPs (G)":"98.3","Params (M)":"28.5"}. This record does not assert independent reproduction.

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