KITTI-360

Method

Image2Maps+SeaBird [I2M+SeaBird]
https://github.com/abhi1kumar/SeaBird

Submitted on 8 May. 2023 21:23 by
Abhinav Kumar (Michigan State University (MSU))

Running time:		0.02 s
Environment:		NVIDIA A100

Method Description:

SeaBird trained on KITTI-360 Train+Val

Parameters:

Trained on 2 classes: Cars and Buildings

Latex Bibtex:

@inproceedings{kumar2024seabird,
title={{SeaBird: Segmentation in Bird's View with Dice Loss Improves Monocular 3D Detection of Large Objects}},
author={Kumar, Abhinav and Guo, Yuliang and Huang, Xinyu and Ren, Liu and Liu, Xiaoming},
booktitle={CVPR},
year={2024}
}

Method

Test Set Average