Method

Point Virtual Transformer V2 [PointVit V2]
[Anonymous Submission]

Submitted on 8 Aug. 2025 19:59 by
[Anonymous Submission]

Running time:.006 s
Environment:1 core @ 2.5 Ghz (Python + C/C++)

Method Description:
This is a single stage transformer architecture
operating directly on voxelized LiDAR points.
PointViT V2 introduces depth based virtual points
with Depth Maps from BP-Net,
integrates self attention with local depthwise
convolutions, and employs a multiscale voxel
feature encoder.
Parameters:
Num_Heads = 4
Latex Bibtex:

Detailed Results

Object detection and orientation estimation results. Results for object detection are given in terms of average precision (AP) and results for joint object detection and orientation estimation are provided in terms of average orientation similarity (AOS).


Benchmark Easy Moderate Hard
Car (Detection) 97.04 % 96.56 % 88.97 %
Car (Orientation) 97.04 % 96.50 % 88.88 %
Car (3D Detection) 89.81 % 80.54 % 74.96 %
Car (Bird's Eye View) 93.59 % 89.67 % 82.12 %
This table as LaTeX

2D object detection results.
This figure as: png eps txt gnuplot



Orientation estimation results.
This figure as: png eps txt gnuplot



3D object detection results.
This figure as: png eps txt gnuplot



Bird's eye view results.
This figure as: png eps txt gnuplot




eXTReMe Tracker