# Instance Segmentation on VOC 2012 (val) - Detailed AP^r Thresholds

Source: [Instance Segmentation on VOC 2012 (val) - Detailed AP^r Thresholds](<https://www.sota2.com/research/sota/instance-segmentation-on-voc-2012-val-detailed-ap-r-thresholds>)

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

Data updated: 2026-02-26T05:23:30.393Z

Evaluation of Instance Segmentation performance on the validation split of PASCAL VOC 2012, reporting Average Precision at specific IoU thresholds (0.5 through 0.9) and volume-based AP^r.

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 | Instance Segmentation |
| Dataset (including any stated split) | VOC 2012 (val) |
| Metric | AP^r @ IoU=0.5 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 13 of 13 |
| 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

- [AP^r @ IoU=0.5](<https://www.sota2.com/research/sota/instance-segmentation-on-voc-2012-val-detailed-ap-r-thresholds.md?metric=AP%5Er%20%40%20IoU%3D0.5>): Higher is better; unit: PERCENT.

- [AP^r @ IoU=0.6](<https://www.sota2.com/research/sota/instance-segmentation-on-voc-2012-val-detailed-ap-r-thresholds.md?metric=AP%5Er%20%40%20IoU%3D0.6>): Higher is better; unit: PERCENT.

- [AP^r @ IoU=0.7](<https://www.sota2.com/research/sota/instance-segmentation-on-voc-2012-val-detailed-ap-r-thresholds.md?metric=AP%5Er%20%40%20IoU%3D0.7>): Higher is better; unit: PERCENT.

- [AP^r @ IoU=0.8](<https://www.sota2.com/research/sota/instance-segmentation-on-voc-2012-val-detailed-ap-r-thresholds.md?metric=AP%5Er%20%40%20IoU%3D0.8>): Higher is better; unit: PERCENT.

- [AP^r @ IoU=0.9](<https://www.sota2.com/research/sota/instance-segmentation-on-voc-2012-val-detailed-ap-r-thresholds.md?metric=AP%5Er%20%40%20IoU%3D0.9>): Higher is better; unit: PERCENT.

- [AP^r (Volume)](<https://www.sota2.com/research/sota/instance-segmentation-on-voc-2012-val-detailed-ap-r-thresholds.md?metric=AP%5Er%20(Volume)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | AP^r @ IoU=0.5 | Result date | Report paper |
| --- | --- | --- | --- |
| Weakly-supervised Panoptic Segmentation | 63.9 | Not recorded | 68fa654b1e209131cea095a8 |
| Weakly-supervised Panoptic Segmentation | 63.6 | Not recorded | 68fa654b1e209131cea095a8 |
| MPA | 62.1 | Not recorded | 68fa654b1e209131cea095a8 |
| Arnab et al. | 61.7 | Not recorded | 68fa654b1e209131cea095a8 |
| SGN | 61.4 | Not recorded | 68fa654b1e209131cea095a8 |
| Weakly-supervised Panoptic Segmentation | 60.9 | Not recorded | 68fa654b1e209131cea095a8 |
| Weakly-supervised Panoptic Segmentation | 60.5 | Not recorded | 68fa654b1e209131cea095a8 |
| PFN | 58.7 | Not recorded | 68fa654b1e209131cea095a8 |
| Arnab et al. | 58.3 | Not recorded | 68fa654b1e209131cea095a8 |
| SDI | 46.4 | Not recorded | 68fa654b1e209131cea095a8 |
| Chen et al. | 46.3 | Not recorded | 68fa654b1e209131cea095a8 |
| SDI | 44.8 | Not recorded | 68fa654b1e209131cea095a8 |
| SDS | 43.8 | Not recorded | 68fa654b1e209131cea095a8 |

### Weakly-supervised Panoptic Segmentation

AP^r @ IoU=0.5: 63.9. Result ID: 698ca47f3ce6cd2163e10cb7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Implementation](<https://github.com/qizhuli/Weakly-Supervised-Panoptic-Segmentation>)

Recorded evaluation configuration: {"supervision":"Fully supervised","training\_data":"with COCO"}. This record does not assert independent reproduction.

### Weakly-supervised Panoptic Segmentation

AP^r @ IoU=0.5: 63.6. Result ID: 698ca47f3ce6cd2163e10cb0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Implementation](<https://github.com/qizhuli/Weakly-Supervised-Panoptic-Segmentation>)

Recorded evaluation configuration: {"supervision":"Fully supervised","training\_data":"without COCO"}. This record does not assert independent reproduction.

### MPA

AP^r @ IoU=0.5: 62.1. Result ID: 698ca47f3ce6cd2163e10cb4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

Recorded evaluation configuration: {"supervision":"Fully supervised","training\_data":"with COCO"}. This record does not assert independent reproduction.

### Arnab et al.

AP^r @ IoU=0.5: 61.7. Result ID: 698ca47f3ce6cd2163e10cb5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63381e209131ce9fd895.md>)

Recorded evaluation configuration: {"supervision":"Fully supervised","training\_data":"with COCO","citation":"\[16\]"}. This record does not assert independent reproduction.

### SGN

AP^r @ IoU=0.5: 61.4. Result ID: 698ca47f3ce6cd2163e10cb6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

Recorded evaluation configuration: {"supervision":"Fully supervised","training\_data":"with COCO"}. This record does not assert independent reproduction.

### Weakly-supervised Panoptic Segmentation

AP^r @ IoU=0.5: 60.9. Result ID: 698ca47f3ce6cd2163e10cb2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Implementation](<https://github.com/qizhuli/Weakly-Supervised-Panoptic-Segmentation>)

Recorded evaluation configuration: {"supervision":"Weakly supervised","training\_data":"with COCO"}. This record does not assert independent reproduction.

### Weakly-supervised Panoptic Segmentation

AP^r @ IoU=0.5: 60.5. Result ID: 698ca47f3ce6cd2163e10cac.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Implementation](<https://github.com/qizhuli/Weakly-Supervised-Panoptic-Segmentation>)

Recorded evaluation configuration: {"supervision":"Weakly supervised","training\_data":"without COCO"}. This record does not assert independent reproduction.

### PFN

AP^r @ IoU=0.5: 58.7. Result ID: 698ca47f3ce6cd2163e10caf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62121e209131ce9f7476.md>)

Recorded evaluation configuration: {"supervision":"Fully supervised","training\_data":"without COCO"}. This record does not assert independent reproduction.

### Arnab et al.

AP^r @ IoU=0.5: 58.3. Result ID: 698ca47f3ce6cd2163e10cb3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63381e209131ce9fd895.md>)

Recorded evaluation configuration: {"supervision":"Fully supervised","training\_data":"with COCO","citation":"\[17\]"}. This record does not assert independent reproduction.

### SDI

AP^r @ IoU=0.5: 46.4. Result ID: 698ca47f3ce6cd2163e10cb1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63001e209131ce9fc38e.md>)

Recorded evaluation configuration: {"supervision":"Weakly supervised","training\_data":"with COCO"}. This record does not assert independent reproduction.

### Chen et al.

AP^r @ IoU=0.5: 46.3. Result ID: 698ca47f3ce6cd2163e10cae.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

Recorded evaluation configuration: {"supervision":"Fully supervised","training\_data":"without COCO"}. This record does not assert independent reproduction.

### SDI

AP^r @ IoU=0.5: 44.8. Result ID: 698ca47f3ce6cd2163e10cab.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63001e209131ce9fc38e.md>)

Recorded evaluation configuration: {"supervision":"Weakly supervised","training\_data":"without COCO"}. This record does not assert independent reproduction.

### SDS

AP^r @ IoU=0.5: 43.8. Result ID: 698ca47f3ce6cd2163e10cad.

[Reporting paper](<https://www.sota2.com/research/paper/68fa654b1e209131cea095a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa616a1e209131ce9f4432.md>)

[Implementation](<http://www.eecs.berkeley.edu/Research/Projects/CS/vision/shape/sds>)

Recorded evaluation configuration: {"supervision":"Fully supervised","training\_data":"without COCO"}. This record does not assert independent reproduction.

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