MFSeg: Efficient Multi-frame 3D Semantic Segmentation

TL;DR

MFSeg is a novel multi-frame 3D semantic segmentation framework that efficiently aggregates point cloud features, reduces computational costs, and maintains high accuracy, outperforming existing methods on major datasets.

Contribution

Introduces a lightweight, feature-level aggregation method with a simple MLP decoder, improving efficiency without sacrificing accuracy in 3D segmentation.

Findings

Outperforms existing methods on nuScenes and Waymo datasets

Reduces computational overhead compared to prior approaches

Maintains high segmentation accuracy with a lightweight design

Abstract

We propose MFSeg, an efficient multi-frame 3D semantic segmentation framework. By aggregating point cloud sequences at the feature level and regularizing the feature extraction and aggregation process, MFSeg reduces computational overhead while maintaining high accuracy. Moreover, by employing a lightweight MLP-based point decoder, our method eliminates the need to upsample redundant points from past frames. Experiments on the nuScenes and Waymo datasets show that MFSeg outperforms existing methods, demonstrating its effectiveness and efficiency.