# 3D Multi-Person Pose Estimation on Human3.6M, Shelf, Campus, and MVOR Averaged Generalization

Source: [3D Multi-Person Pose Estimation on Human3.6M, Shelf, Campus, and MVOR Averaged Generalization](<https://www.sota2.com/research/sota/3d-multi-person-pose-estimation-on-human3-6m-shelf-campus-and-mvor-averaged-generalization>)

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

Data updated: 2026-02-26T05:54:58.017Z

Evaluation of 3D Multi-Person Pose Estimation performance averaged across generalization tests involving Human3.6M, Shelf, Campus, and MVOR datasets.

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 | 3D Multi-Person Pose Estimation |
| Dataset (including any stated split) | Human3.6M, Shelf, Campus, and MVOR Averaged Generalization |
| Metric | PCP |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [PCP](<https://www.sota2.com/research/sota/3d-multi-person-pose-estimation-on-human3-6m-shelf-campus-and-mvor-averaged-generalization.md?metric=PCP>): Higher is better; unit: PERCENT.

- [PCK@100](<https://www.sota2.com/research/sota/3d-multi-person-pose-estimation-on-human3-6m-shelf-campus-and-mvor-averaged-generalization.md?metric=PCK%40100>): Higher is better; unit: PERCENT.

- [PCK@500](<https://www.sota2.com/research/sota/3d-multi-person-pose-estimation-on-human3-6m-shelf-campus-and-mvor-averaged-generalization.md?metric=PCK%40500>): Higher is better; unit: PERCENT.

- [MPJPE](<https://www.sota2.com/research/sota/3d-multi-person-pose-estimation-on-human3-6m-shelf-campus-and-mvor-averaged-generalization.md?metric=MPJPE>): Lower is better; unit: SCALAR.

- [Recall@100](<https://www.sota2.com/research/sota/3d-multi-person-pose-estimation-on-human3-6m-shelf-campus-and-mvor-averaged-generalization.md?metric=Recall%40100>): Higher is better; unit: PERCENT.

- [Recall@500](<https://www.sota2.com/research/sota/3d-multi-person-pose-estimation-on-human3-6m-shelf-campus-and-mvor-averaged-generalization.md?metric=Recall%40500>): Higher is better; unit: PERCENT.

- [Invalid Rate](<https://www.sota2.com/research/sota/3d-multi-person-pose-estimation-on-human3-6m-shelf-campus-and-mvor-averaged-generalization.md?metric=Invalid%20Rate>): Lower is better; unit: PERCENT.

- [F1 Score](<https://www.sota2.com/research/sota/3d-multi-person-pose-estimation-on-human3-6m-shelf-campus-and-mvor-averaged-generalization.md?metric=F1%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | PCP | Result date | Report paper |
| --- | --- | --- | --- |
| VoxelKeypointFusion | 85.3 | Not recorded | 68fa764b1e209131cea6c362 |
| mvpose | 77.6 | Not recorded | 68fa764b1e209131cea6c362 |
| PartAwarePose | 73.9 | Not recorded | 68fa764b1e209131cea6c362 |
| Faster-VoxelPose | 73.5 | Not recorded | 68fa764b1e209131cea6c362 |
| VoxelPose | 73.3 | Not recorded | 68fa764b1e209131cea6c362 |
| Faster-VoxelPose | 72.1 | Not recorded | 68fa764b1e209131cea6c362 |
| SelfPose3d | 60.3 | Not recorded | 68fa764b1e209131cea6c362 |
| mv3dpose | 59.6 | Not recorded | 68fa764b1e209131cea6c362 |
| TEMPO | 59.1 | Not recorded | 68fa764b1e209131cea6c362 |
| PRGnet | 47.1 | Not recorded | 68fa764b1e209131cea6c362 |
| VoxelPose | 37.6 | Not recorded | 68fa764b1e209131cea6c362 |
| MvP | 1.8 | Not recorded | 68fa764b1e209131cea6c362 |

### VoxelKeypointFusion

PCP: 85.3. Result ID: 699419183a6c3148491b0843.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Implementation](<https://gitlab.com/Percipiote/>)

Recorded evaluation configuration: {"FPS":8.4}. This record does not assert independent reproduction.

### mvpose

PCP: 77.6. Result ID: 699419183a6c3148491b0840.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa654d1e209131cea09772.md>)

[Implementation](<https://zju-3dv.github.io/mvpose/>)

Recorded evaluation configuration: {"FPS":0.4}. This record does not assert independent reproduction.

### PartAwarePose

PCP: 73.9. Result ID: 699419183a6c3148491b0842.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a3f1e209131cea26ff7.md>)

[Implementation](<https://git.io/JO4KE>)

Recorded evaluation configuration: {"FPS":4.8}. This record does not assert independent reproduction.

### Faster-VoxelPose

PCP: 73.5. Result ID: 699419183a6c3148491b083b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d121e209131cea37746.md>)

Recorded evaluation configuration: {"training\_data":"synthetic","FPS":27.2}. This record does not assert independent reproduction.

### VoxelPose

PCP: 73.3. Result ID: 699419183a6c3148491b0839.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

Recorded evaluation configuration: {"training\_data":"synthetic","FPS":10.9}. This record does not assert independent reproduction.

### Faster-VoxelPose

PCP: 72.1. Result ID: 699419183a6c3148491b083a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d121e209131cea37746.md>)

Recorded evaluation configuration: {"FPS":26.2}. This record does not assert independent reproduction.

### SelfPose3d

PCP: 60.3. Result ID: 699419183a6c3148491b083f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73861e209131cea5cf17.md>)

[Implementation](<https://github.com/CAMMA-public/SelfPose3D>)

Recorded evaluation configuration: {"FPS":8.1}. This record does not assert independent reproduction.

### mv3dpose

PCP: 59.6. Result ID: 699419183a6c3148491b0841.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa692e1e209131cea20a30.md>)

[Implementation](<https://github.com/jutanke/mv3dpose>)

Recorded evaluation configuration: {"FPS":2.5}. This record does not assert independent reproduction.

### TEMPO

PCP: 59.1. Result ID: 699419183a6c3148491b083e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa708f1e209131cea4bcf5.md>)

Recorded evaluation configuration: {"FPS":11.4}. This record does not assert independent reproduction.

### PRGnet

PCP: 47.1. Result ID: 699419183a6c3148491b083d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ab51e209131cea29ec6.md>)

Recorded evaluation configuration: {"FPS":12.1}. This record does not assert independent reproduction.

### VoxelPose

PCP: 37.6. Result ID: 699419183a6c3148491b0838.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

Recorded evaluation configuration: {"FPS":13.9}. This record does not assert independent reproduction.

### MvP

PCP: 1.8. Result ID: 699419183a6c3148491b083c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa764b1e209131cea6c362.md>)

Recorded evaluation configuration: {"FPS":8.9}. This record does not assert independent reproduction.

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