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

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

21 Mar 2024CVPR 2024 1arXiv:2403.14418archive 2025-07-28

Bohao Peng, Xiaoyang Wu, Li Jiang, Yukang Chen, Hengshuang Zhao, Zhuotao Tian, Jiaya Jia

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.

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

Pointcept/Pointcept officialmentioned in papermentioned on GitHubpytorchMIT report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

3D Semantic SegmentationLIDAR Semantic SegmentationSemantic Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
3D Semantic Segmentation ScanNet++ OA-CNN Top-1 IoU 0.470 #5 of 8 Archive leaderboard report
3D Semantic Segmentation ScanNet++ OA-CNN Top-3 IoU 0.726 #5 of 8 Archive leaderboard report
3D Semantic Segmentation ScanNet200 OA-CNNs test mIoU 32.3 #8 of 16 Archive leaderboard report
3D Semantic Segmentation ScanNet200 OA-CNNs val mIoU 33.3 #8 of 16 Archive leaderboard report
3D Semantic Segmentation SemanticKITTI OA-CNNs val mIoU 70.6% #43 of 45 Archive leaderboard report
LIDAR Semantic Segmentation nuScenes OA-CNNs val mIoU 0.789 #34 of 36 Archive leaderboard report
Semantic Segmentation ScanNet OA-CNNs test mIoU 75.6 #14 of 45 Archive leaderboard report
Semantic Segmentation ScanNet OA-CNNs val mIoU 76.1 #14 of 45 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

SPEED

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections