# 3D Object Detection on KITTI tracking Hard (val) Source: [3D Object Detection on KITTI tracking Hard (val)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:23:04.135Z Evaluation of 3D Object Detection performance specifically on the 'Hard' difficulty subset of the KITTI tracking validation set, reporting class-wise Average Precision (AP) for Car, Pedestrian, and Cyclist categories. 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 tracking Hard (val) | | Metric | AP (Car) | | 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 - [AP (Car)](): Higher is better; unit: PERCENT. - [AP (Pedestrian)](): Higher is better; unit: PERCENT. - [AP (Cyclist)](): Higher is better; unit: PERCENT. - [mAP](): Higher is better; unit: PERCENT. ## Recorded results | Method | AP (Car) | Result date | Report paper | | --- | --- | --- | --- | | 3D-FCT | 72.72 | Not recorded | 68fa6aee1e209131cea2b245 | | PV-RCNN + T loss | 71.28 | Not recorded | 68fa6aee1e209131cea2b245 | | PV-RCNN | 70.57 | Not recorded | 68fa6aee1e209131cea2b245 | | SECOND | 67.17 | Not recorded | 68fa6aee1e209131cea2b245 | | PointRCNN | 65.08 | Not recorded | 68fa6aee1e209131cea2b245 | | PointPillar | 49.64 | Not recorded | 68fa6aee1e209131cea2b245 | ### 3D-FCT AP (Car): 72.72. Result ID: 698e4235ffc3901de09c17fb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"backbone":"PV-RCNN"}. This record does not assert independent reproduction. ### PV-RCNN + T loss AP (Car): 71.28. Result ID: 698e4235ffc3901de09c17fa. [Reporting paper]() Recorded evaluation configuration: {"tracking loss":"true"}. This record does not assert independent reproduction. ### PV-RCNN AP (Car): 70.57. Result ID: 698e4235ffc3901de09c17f9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### SECOND AP (Car): 67.17. Result ID: 698e4235ffc3901de09c17f7. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### PointRCNN AP (Car): 65.08. Result ID: 698e4235ffc3901de09c17f8. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### PointPillar AP (Car): 49.64. Result ID: 698e4235ffc3901de09c17f6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.