Method

MonoDSSMs-A [MonoDSSMs-A]


Submitted on 26 Jul. 2024 04:57 by
Đặng Kiệt (TickLab)

Running time:0.02 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
Another version of MonoDSSMs-M with cross-attention
mechanisms in the decoder
Parameters:
The model use 23M parameters.
Latex Bibtex:
@InProceedings{Vu_2024_ACCV,
author = {Vu, Kiet Dang and Tran, Trung
Thai and Nguyen, Duc Dung},
title = {MonoDSSMs: Efficient Monocular 3D
Object Detection with Depth-Aware State Space
Models},
booktitle = {Proceedings of the Asian
Conference on Computer Vision (ACCV)},
month = {December},
year = {2024},
pages = {3883-3900}
}

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) 93.91 % 88.19 % 76.04 %
Car (Orientation) 93.07 % 86.57 % 74.32 %
Car (3D Detection) 21.47 % 14.55 % 11.78 %
Car (Bird's Eye View) 28.84 % 19.54 % 16.30 %
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