# 3D Object Detection on SI3D-DI Hard (test) Source: [3D Object Detection on SI3D-DI Hard (test)]() Collection index: [llms.txt]() Data updated: 2026-06-09T10:24:18.190Z Evaluation of 3D Object Detection performance on the hard difficulty test split of the SI3D-DI dataset, reporting category-specific AP metrics (APBEV, AP3D) and overall mean AP metrics (mAPBEV, mAP3D). 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) | SI3D-DI Hard (test) | | Metric | Vehicle APBEV | | 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 - [Vehicle APBEV](): Higher is better; unit: PERCENT. - [Vehicle AP3D](): Higher is better; unit: PERCENT. - [Articulated Target APBEV](): Higher is better; unit: PERCENT. - [Articulated Target AP3D](): Higher is better; unit: PERCENT. - [Camera Target APBEV](): Higher is better; unit: PERCENT. - [Camera Target AP3D](): Higher is better; unit: PERCENT. - [Canopy APBEV](): Higher is better; unit: PERCENT. - [Canopy AP3D](): Higher is better; unit: PERCENT. - [mAPBEV](): Higher is better; unit: PERCENT. - [mAP3D](): Higher is better; unit: PERCENT. ## Recorded results | Method | Vehicle APBEV | Result date | Report paper | | --- | --- | --- | --- | | CAMF-Det | 67.86 | Not recorded | 6a279e5490966ade919bf9b9 | | TG-ADet | 67.42 | Not recorded | 6a279e5490966ade919bf9b9 | | VoxelNext | 67.23 | Not recorded | 6a279e5490966ade919bf9b9 | | SAFDNet | 61.52 | Not recorded | 6a279e5490966ade919bf9b9 | | SSLFusion | 57.47 | Not recorded | 6a279e5490966ade919bf9b9 | | Voxelmamba | 56.43 | Not recorded | 6a279e5490966ade919bf9b9 | | HEDNet | 55.93 | Not recorded | 6a279e5490966ade919bf9b9 | | IS-Fusion | 55.28 | Not recorded | 6a279e5490966ade919bf9b9 | | AutoAlignv2 | 52.62 | Not recorded | 6a279e5490966ade919bf9b9 | | TransFusion-L | 49.86 | Not recorded | 6a279e5490966ade919bf9b9 | | BEVFusion | 49.85 | Not recorded | 6a279e5490966ade919bf9b9 | | BEVDilation | 32.52 | Not recorded | 6a279e5490966ade919bf9b9 | ### CAMF-Det Vehicle APBEV: 67.86. Result ID: 6a27e9cefbc373592c25560a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modal":"L+C","Reference":"Ours"}. This record does not assert independent reproduction. ### TG-ADet Vehicle APBEV: 67.42. Result ID: 6a27e9cefbc373592c255604. [Reporting paper]() Recorded evaluation configuration: {"Modal":"L","Reference":"TGRS 2025"}. This record does not assert independent reproduction. ### VoxelNext Vehicle APBEV: 67.23. Result ID: 6a27e9cefbc373592c255601. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modal":"L","Reference":"CVPR 2023"}. This record does not assert independent reproduction. ### SAFDNet Vehicle APBEV: 61.52. Result ID: 6a27e9cefbc373592c255603. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modal":"L","Reference":"CVPR 2024"}. This record does not assert independent reproduction. ### SSLFusion Vehicle APBEV: 57.47. Result ID: 6a27e9cefbc373592c255608. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modal":"L+C","Reference":"AAAI 2025"}. This record does not assert independent reproduction. ### Voxelmamba Vehicle APBEV: 56.43. Result ID: 6a27e9cefbc373592c255602. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modal":"L","Reference":"NeurIPS 2024"}. This record does not assert independent reproduction. ### HEDNet Vehicle APBEV: 55.93. Result ID: 6a27e9cefbc373592c255600. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modal":"L","Reference":"NeurIPS 2023"}. This record does not assert independent reproduction. ### IS-Fusion Vehicle APBEV: 55.28. Result ID: 6a27e9cefbc373592c255607. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modal":"L+C","Reference":"CVPR 2024"}. This record does not assert independent reproduction. ### AutoAlignv2 Vehicle APBEV: 52.62. Result ID: 6a27e9cefbc373592c255605. [Reporting paper]() Recorded evaluation configuration: {"Modal":"L+C","Reference":"ECCV 2022"}. This record does not assert independent reproduction. ### TransFusion-L Vehicle APBEV: 49.86. Result ID: 6a27e9cefbc373592c2555ff. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modal":"L","Reference":"CVPR 2022"}. This record does not assert independent reproduction. ### BEVFusion Vehicle APBEV: 49.85. Result ID: 6a27e9cefbc373592c255606. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modal":"L+C","Reference":"ICRA 2023"}. This record does not assert independent reproduction. ### BEVDilation Vehicle APBEV: 32.52. Result ID: 6a27e9cefbc373592c255609. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modal":"L+C","Reference":"AAAI 2026"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.