Method

Mamba-fusion [MFusion]
[Anonymous Submission]

Submitted on 30 Mar. 2026 18:02 by
[Anonymous Submission]

Running time:0.11 s
Environment:1 core @ 2.5 Ghz (Python)

Method Description:
Our research focuses on multi-modal 3D object
detection by integrating LiDAR point clouds and
RGB images for autonomous driving. We propose a
novel fusion framework, Voxel-Mamba-Fusion, which
leverages state space models (SSMs) to model
cross-modal interactions in a unified sequential
manner.
Parameters:
Batch_size=48
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) 96.13 % 92.65 % 89.91 %
Car (Orientation) 96.06 % 92.38 % 89.50 %
Car (3D Detection) 86.92 % 78.50 % 73.81 %
Car (Bird's Eye View) 91.66 % 88.01 % 83.39 %
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