OA-CNNs: Omni-Adaptive Sparse CNNs for 3D Semantic Segmentation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Peng, Bohao, Wu, Xiaoyang, Jiang, Li, Chen, Yukang, Zhao, Hengshuang, Tian, Zhuotao, Jia, Jiaya
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909144949194752
author Peng, Bohao
Wu, Xiaoyang
Jiang, Li
Chen, Yukang
Zhao, Hengshuang
Tian, Zhuotao
Jia, Jiaya
author_facet Peng, Bohao
Wu, Xiaoyang
Jiang, Li
Chen, Yukang
Zhao, Hengshuang
Tian, Zhuotao
Jia, Jiaya
contents The booming of 3D recognition in the 2020s began with the introduction of point cloud transformers. They quickly overwhelmed sparse CNNs and became state-of-the-art models, especially in 3D semantic segmentation. However, sparse CNNs are still valuable networks, due to their efficiency treasure, and ease of application. In this work, we reexamine the design distinctions and test the limits of what a sparse CNN can achieve. We discover that the key credit to the performance difference is adaptivity. Specifically, we propose two key components, i.e., adaptive receptive fields (spatially) and adaptive relation, to bridge the gap. This exploration led to the creation of Omni-Adaptive 3D CNNs (OA-CNNs), a family of networks that integrates a lightweight module to greatly enhance the adaptivity of sparse CNNs at minimal computational cost. Without any self-attention modules, OA-CNNs favorably surpass point transformers in terms of accuracy in both indoor and outdoor scenes, with much less latency and memory cost. Notably, it achieves 76.1%, 78.9%, and 70.6% mIoU on ScanNet v2, nuScenes, and SemanticKITTI validation benchmarks respectively, while maintaining at most 5x better speed than transformer counterparts. This revelation highlights the potential of pure sparse CNNs to outperform transformer-related networks.
format Preprint
id arxiv_https___arxiv_org_abs_2403_14418
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle OA-CNNs: Omni-Adaptive Sparse CNNs for 3D Semantic Segmentation
Peng, Bohao
Wu, Xiaoyang
Jiang, Li
Chen, Yukang
Zhao, Hengshuang
Tian, Zhuotao
Jia, Jiaya
Computer Vision and Pattern Recognition
The booming of 3D recognition in the 2020s began with the introduction of point cloud transformers. They quickly overwhelmed sparse CNNs and became state-of-the-art models, especially in 3D semantic segmentation. However, sparse CNNs are still valuable networks, due to their efficiency treasure, and ease of application. In this work, we reexamine the design distinctions and test the limits of what a sparse CNN can achieve. We discover that the key credit to the performance difference is adaptivity. Specifically, we propose two key components, i.e., adaptive receptive fields (spatially) and adaptive relation, to bridge the gap. This exploration led to the creation of Omni-Adaptive 3D CNNs (OA-CNNs), a family of networks that integrates a lightweight module to greatly enhance the adaptivity of sparse CNNs at minimal computational cost. Without any self-attention modules, OA-CNNs favorably surpass point transformers in terms of accuracy in both indoor and outdoor scenes, with much less latency and memory cost. Notably, it achieves 76.1%, 78.9%, and 70.6% mIoU on ScanNet v2, nuScenes, and SemanticKITTI validation benchmarks respectively, while maintaining at most 5x better speed than transformer counterparts. This revelation highlights the potential of pure sparse CNNs to outperform transformer-related networks.
title OA-CNNs: Omni-Adaptive Sparse CNNs for 3D Semantic Segmentation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.14418