In this technical report, we detail our first-place solution for the 2024 Waymo Open Dataset Challenge's semantic segmentation track. We significantly enhanced the performance of Point Transformer V3 on the Waymo benchmark by implementing cutting-edge, plug-and-play training and inference technologies. Notably, our advanced version, Point Transformer V3 Extreme, leverages multi-frame training and a no-clipping-point policy, achieving substantial gains over the original PTv3 performance. Additionally, employing a straightforward model ensemble strategy further boosted our results. This approach secured us the top position on the Waymo Open Dataset semantic segmentation leaderboard, markedly outperforming other entries.
@article{arxiv.2407.15282,
title = {Point Transformer V3 Extreme: 1st Place Solution for 2024 Waymo Open Dataset Challenge in Semantic Segmentation},
author = {Xiaoyang Wu and Xiang Xu and Lingdong Kong and Liang Pan and Ziwei Liu and Tong He and Wanli Ouyang and Hengshuang Zhao},
journal= {arXiv preprint arXiv:2407.15282},
year = {2024}
}
Comments
1st Place Solution for 2024 Waymo Open Dataset Challenge in Semantic Segmentation