# 3D Object Detection on Waymo Open Dataset (WOD) (val) (Granular 3D AP)

Source: [3D Object Detection on Waymo Open Dataset (WOD) (val) (Granular 3D AP)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap>)

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

Data updated: 2026-02-26T06:08:11.768Z

Evaluation of 3D Object Detection performance on the validation split of the Waymo Open Dataset (WOD). Metrics reported are granular 3D Average Precision (AP) scores for Vehicle, Pedestrian, and Cyclist classes, segmented by L1/L2 difficulty levels and specific IoU thresholds (0.3, 0.5, and 0.7).

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 Object Detection |
| Dataset (including any stated split) | Waymo Open Dataset (WOD) (val) |
| Metric | Vehicle 3D AP L1 (IoU=0.5) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [Vehicle 3D AP L1 (IoU=0.5)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Vehicle%203D%20AP%20L1%20(IoU%3D0.5)>): Higher is better; unit: PERCENT.

- [Vehicle 3D AP L1 (IoU=0.7)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Vehicle%203D%20AP%20L1%20(IoU%3D0.7)>): Higher is better; unit: PERCENT.

- [Vehicle 3D AP L2 (IoU=0.5)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Vehicle%203D%20AP%20L2%20(IoU%3D0.5)>): Higher is better; unit: PERCENT.

- [Vehicle 3D AP L2 (IoU=0.7)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Vehicle%203D%20AP%20L2%20(IoU%3D0.7)>): Higher is better; unit: PERCENT.

- [Pedestrian 3D AP L1 (IoU=0.3)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Pedestrian%203D%20AP%20L1%20(IoU%3D0.3)>): Higher is better; unit: PERCENT.

- [Pedestrian 3D AP L1 (IoU=0.5)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Pedestrian%203D%20AP%20L1%20(IoU%3D0.5)>): Higher is better; unit: PERCENT.

- [Pedestrian 3D AP L2 (IoU=0.3)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Pedestrian%203D%20AP%20L2%20(IoU%3D0.3)>): Higher is better; unit: PERCENT.

- [Pedestrian 3D AP L2 (IoU=0.5)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Pedestrian%203D%20AP%20L2%20(IoU%3D0.5)>): Higher is better; unit: PERCENT.

- [Cyclist 3D AP L1 (IoU=0.3)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Cyclist%203D%20AP%20L1%20(IoU%3D0.3)>): Higher is better; unit: PERCENT.

- [Cyclist 3D AP L1 (IoU=0.5)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Cyclist%203D%20AP%20L1%20(IoU%3D0.5)>): Higher is better; unit: PERCENT.

- [Cyclist 3D AP L2 (IoU=0.3)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Cyclist%203D%20AP%20L2%20(IoU%3D0.3)>): Higher is better; unit: PERCENT.

- [Cyclist 3D AP L2 (IoU=0.5)](<https://www.sota2.com/research/sota/3d-object-detection-on-waymo-open-dataset-wod-val-granular-3d-ap.md?metric=Cyclist%203D%20AP%20L2%20(IoU%3D0.5)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Vehicle 3D AP L1 (IoU=0.5) | Result date | Report paper |
| --- | --- | --- | --- |
| CPD | 57.79 | Not recorded | 68fa73e41e209131cea5eb63 |
| Proto-vanilla | 35.22 | Not recorded | 68fa73e41e209131cea5eb63 |
| OYSTER | 30.48 | Not recorded | 68fa73e41e209131cea5eb63 |
| MODEST | 18.51 | Not recorded | 68fa73e41e209131cea5eb63 |
| DBSCAN init-train | 17.36 | Not recorded | 68fa73e41e209131cea5eb63 |
| DBSCAN | 2.32 | Not recorded | 68fa73e41e209131cea5eb63 |

### CPD

Vehicle 3D AP L1 (IoU=0.5): 57.79. Result ID: 699324493a6c31484914006c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e41e209131cea5eb63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73e41e209131cea5eb63.md>)

[Implementation](<https://github.com/hailanyi/CPD>)

Recorded evaluation configuration: {"supervision":"unsupervised","backbone":"CenterPoint-like"}. This record does not assert independent reproduction.

### Proto-vanilla

Vehicle 3D AP L1 (IoU=0.5): 35.22. Result ID: 699324493a6c31484914006b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e41e209131cea5eb63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa636e1e209131ce9febd3.md>)

Recorded evaluation configuration: {"supervision":"unsupervised"}. This record does not assert independent reproduction.

### OYSTER

Vehicle 3D AP L1 (IoU=0.5): 30.48. Result ID: 699324493a6c31484914006a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e41e209131cea5eb63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa710e1e209131cea4ef13.md>)

Recorded evaluation configuration: {"supervision":"unsupervised"}. This record does not assert independent reproduction.

### MODEST

Vehicle 3D AP L1 (IoU=0.5): 18.51. Result ID: 699324493a6c314849140069.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e41e209131cea5eb63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c181e209131cea32213.md>)

[Implementation](<https://github.com/YurongYou/MODEST>)

Recorded evaluation configuration: {"supervision":"unsupervised"}. This record does not assert independent reproduction.

### DBSCAN init-train

Vehicle 3D AP L1 (IoU=0.5): 17.36. Result ID: 699324493a6c314849140068.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e41e209131cea5eb63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c181e209131cea32213.md>)

[Implementation](<https://github.com/YurongYou/MODEST>)

Recorded evaluation configuration: {"supervision":"unsupervised"}. This record does not assert independent reproduction.

### DBSCAN

Vehicle 3D AP L1 (IoU=0.5): 2.32. Result ID: 699324493a6c314849140067.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e41e209131cea5eb63.md>)

Recorded evaluation configuration: {"supervision":"unsupervised"}. This record does not assert independent reproduction.

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