KITTI-360

Method

DEVIANT: Depth Equivariant Network for Moncoular 3D Object Detection [DEVIANT]
https://github.com/abhi1kumar/DEVIANT

Submitted on 17 Apr. 2023 23:26 by
Andrew Hou (Michigan State University)

Running time:		0.04 s
Environment:		NVIDIA A100

Method Description:

DEVIANT (ECCV 2022) trained on KITTI-360 Train+Val

Parameters:

sesn_scales: [0.83, 1.0 ]
2 classes: Cars and Buildings

Latex Bibtex:

@inproceedings{kumar2022deviant,
title={{DEVIANT: Depth Equivariant Network for Monocular 3D Object Detection}},
author={Kumar, Abhinav and Brazil, Garrick and Corona, Enrique and Parchami, Armin and Liu, Xiaoming},
booktitle={ECCV},
year={2022}
}

Method

Test Set Average