Mamba3D: Enhancing Local Features for 3D Point Cloud Analysis via State Space Model

TL;DR

Mamba3D introduces a novel state space model tailored for 3D point cloud analysis, enhancing local feature extraction and outperforming Transformer-based models in accuracy and efficiency.

Contribution

The paper presents Mamba3D, a new state space model with Local Norm Pooling and bidirectional SSM for improved local and global feature extraction in point clouds.

Findings

Achieves state-of-the-art accuracy on ScanObjectNN and ModelNet40.

Outperforms Transformer-based models in multiple point cloud tasks.

Maintains linear complexity, enabling scalable analysis.

Abstract

Existing Transformer-based models for point cloud analysis suffer from quadratic complexity, leading to compromised point cloud resolution and information loss. In contrast, the newly proposed Mamba model, based on state space models (SSM), outperforms Transformer in multiple areas with only linear complexity. However, the straightforward adoption of Mamba does not achieve satisfactory performance on point cloud tasks. In this work, we present Mamba3D, a state space model tailored for point cloud learning to enhance local feature extraction, achieving superior performance, high efficiency, and scalability potential. Specifically, we propose a simple yet effective Local Norm Pooling (LNP) block to extract local geometric features. Additionally, to obtain better global features, we introduce a bidirectional SSM (bi-SSM) with both a token forward SSM and a novel backward SSM that operates on the feature channel. Extensive experimental results show that Mamba3D surpasses Transformer-based counterparts and concurrent works in multiple tasks, with or without pre-training. Notably, Mamba3D achieves multiple SoTA, including an overall accuracy of 92.6% (train from scratch) on the ScanObjectNN and 95.1% (with single-modal pre-training) on the ModelNet40 classification task, with only linear complexity. Our code and weights are available at https://github.com/xhanxu/Mamba3D.