Pointpillars_nuscenes - 2.0 English

Vitis AI Library User Guide (UG1354)

Document ID
UG1354
Release Date
2022-01-20
Version
2.0 English

PointPillars is an efficient network for real-time 3D object detection on the point cloud. Trained on the nuScenes dataset, this model gives 3D bounding boxes and speed prediction for ten classes (including some kinds of vehicles, pedestrians, barriers, and traffic cones) in the surround-view range. With multisweep point clouds as input, PointPillars can achieve higher accuracy of 3D object detection and speed estimation at the cost of increasing the complexity of the pre-processing part.

Figure 1. Pointpillars_nuscenes Example

The following table lists the Pointpillars_nuscenes models supported by the Vitis AI library.

Table 1. Pointpillars_nuscenes Models
No Model Name Framework
1 pointpillars_nuscenes_40000_64_0_pt PyTorch
2 pointpillars_nuscenes_40000_64_1_pt PyTorch