# 3D Object Detection on KITTI official leaderboard (new evaluation metric) (test)

Source: [3D Object Detection on KITTI official leaderboard (new evaluation metric) (test)](<https://www.sota2.com/research/sota/3d-object-detection-on-kitti-official-leaderboard-new-evaluation-metric-test>)

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

Data updated: 2026-02-26T05:23:05.198Z

Evaluation of 3D Object Detection performance on the KITTI official test set using a newly defined evaluation metric, reporting Average Precision across Easy, Moderate, and Hard difficulty levels.

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) | KITTI official leaderboard (new evaluation metric) (test) |
| Metric | AP (Easy) |
| Unit | PERCENT |
| Direction | Higher 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

- [AP (Easy)](<https://www.sota2.com/research/sota/3d-object-detection-on-kitti-official-leaderboard-new-evaluation-metric-test.md?metric=AP%20(Easy)>): Higher is better; unit: PERCENT.

- [AP (Moderate)](<https://www.sota2.com/research/sota/3d-object-detection-on-kitti-official-leaderboard-new-evaluation-metric-test.md?metric=AP%20(Moderate)>): Higher is better; unit: PERCENT.

- [AP (Hard)](<https://www.sota2.com/research/sota/3d-object-detection-on-kitti-official-leaderboard-new-evaluation-metric-test.md?metric=AP%20(Hard)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | AP (Easy) | Result date | Report paper |
| --- | --- | --- | --- |
| DSGN | 73.5 | Not recorded | 68fa67521e209131cea15711 |
| MV3D (LiDAR) | 68.35 | Not recorded | 68fa67521e209131cea15711 |
| PL++: P-RCNN\* | 61.11 | Not recorded | 68fa67521e209131cea15711 |
| PL: AVOD\* | 54.53 | Not recorded | 68fa67521e209131cea15711 |
| Stereo R-CNN\* | 47.58 | Not recorded | 68fa67521e209131cea15711 |
| AM3D | 16.5 | Not recorded | 68fa67521e209131cea15711 |
| M3D-RPN | 14.76 | Not recorded | 68fa67521e209131cea15711 |
| MonoGRNet | 9.61 | Not recorded | 68fa67521e209131cea15711 |
| OFT-Net | 1.61 | Not recorded | 68fa67521e209131cea15711 |

### DSGN

AP (Easy): 73.5. Result ID: 698d2ab78b517eb4331f7f20.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Implementation](<https://github.com/chenyilun95/DSGN>)

Recorded evaluation configuration: {"Modality":"Stereo"}. This record does not assert independent reproduction.

### MV3D (LiDAR)

AP (Easy): 68.35. Result ID: 698d2ab78b517eb4331f7f18.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa636e1e209131ce9fec6f.md>)

Recorded evaluation configuration: {"Modality":"LiDAR"}. This record does not assert independent reproduction.

### PL++: P-RCNN\*

AP (Easy): 61.11. Result ID: 698d2ab78b517eb4331f7f1f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67041e209131cea13c85.md>)

[Implementation](<https://github.com/mileyan/Pseudo_Lidar_V2>)

Recorded evaluation configuration: {"Modality":"Stereo","Pre-training":"Scene Flow"}. This record does not assert independent reproduction.

### PL: AVOD\*

AP (Easy): 54.53. Result ID: 698d2ab78b517eb4331f7f1e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67171e209131cea140c9.md>)

[Implementation](<https://github.com/mileyan/pseudo_lidar>)

Recorded evaluation configuration: {"Modality":"Stereo","Pre-training":"Scene Flow"}. This record does not assert independent reproduction.

### Stereo R-CNN\*

AP (Easy): 47.58. Result ID: 698d2ab78b517eb4331f7f1d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65a31e209131cea0b802.md>)

[Implementation](<https://github.com/HKUST-Aerial-Robotics/Stereo-RCNN>)

Recorded evaluation configuration: {"Modality":"Stereo","Pre-training":"ImageNet"}. This record does not assert independent reproduction.

### AM3D

AP (Easy): 16.5. Result ID: 698d2ab78b517eb4331f7f1c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69981e209131cea22ed9.md>)

Recorded evaluation configuration: {"Modality":"Mono"}. This record does not assert independent reproduction.

### M3D-RPN

AP (Easy): 14.76. Result ID: 698d2ab78b517eb4331f7f1b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa663a1e209131cea0ef88.md>)

[Implementation](<http://cvlab.cse.msu.edu/project-m3d-rpn.html>)

Recorded evaluation configuration: {"Modality":"Mono"}. This record does not assert independent reproduction.

### MonoGRNet

AP (Easy): 9.61. Result ID: 698d2ab78b517eb4331f7f1a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67581e209131cea15b29.md>)

[Implementation](<https://sites.google.com/view/monogrnet>)

Recorded evaluation configuration: {"Modality":"Mono"}. This record does not assert independent reproduction.

### OFT-Net

AP (Easy): 1.61. Result ID: 698d2ab78b517eb4331f7f19.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67521e209131cea15711.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65141e209131cea080e2.md>)

Recorded evaluation configuration: {"Modality":"Mono"}. This record does not assert independent reproduction.

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