MAMBA4D: Efficient Long-Sequence Point Cloud Video Understanding with Disentangled Spatial-Temporal State Space Models
This work addresses the problem of inefficient long-sequence point cloud video understanding for applications like robotics and autonomous systems, offering a novel and more efficient backbone.
The paper tackles the challenge of designing an efficient 4D backbone for point cloud video understanding by proposing MAMBA4D, a method based on State Space Models that disentangles spatial and temporal processing, achieving competitive performance on action recognition, segmentation, and semantic segmentation datasets with significant efficiency gains, including 87.5% GPU memory reduction and 5.36 times speed-up for long sequences.
Point cloud videos can faithfully capture real-world spatial geometries and temporal dynamics, which are essential for enabling intelligent agents to understand the dynamically changing world. However, designing an effective 4D backbone remains challenging, mainly due to the irregular and unordered distribution of points and temporal inconsistencies across frames. Also, recent transformer-based 4D backbones commonly suffer from large computational costs due to their quadratic complexity, particularly for long video sequences. To address these challenges, we propose a novel point cloud video understanding backbone purely based on the State Space Models (SSMs). Specifically, we first disentangle space and time in 4D video sequences and then establish the spatio-temporal correlation with our designed Mamba blocks. The Intra-frame Spatial Mamba module is developed to encode locally similar geometric structures within a certain temporal stride. Subsequently, locally correlated tokens are delivered to the Inter-frame Temporal Mamba module, which integrates long-term point features across the entire video with linear complexity. Our proposed Mamba4d achieves competitive performance on the MSR-Action3D action recognition (+10.4% accuracy), HOI4D action segmentation (+0.7 F1 Score), and Synthia4D semantic segmentation (+0.19 mIoU) datasets. Especially, for long video sequences, our method has a significant efficiency improvement with 87.5% GPU memory reduction and 5.36 times speed-up. Codes will be released at https://github.com/IRMVLab/Mamba4D.