DAPointMamba: Domain Adaptive Point Mamba for Point Cloud Completion

TL;DR

Domain Adaptive Point Cloud Completion faces distribution gaps between labeled sources and unlabeled targets. DAPointMamba introduces a Mamba-based framework with three cross-domain modules—Cross-Domain Patch-Level Scanning (CDPS), Cross-Domain Spatial SSM Alignment (CDSA), and Cross-Domain Channel SSM Alignment (CDCA)—to achieve robust local and global alignment with linear computational complexity. Across synthetic and real-world benchmarks, it surpasses state-of-the-art methods while reducing parameters, FLOPs, and latency, demonstrating strong cross-domain transferability and efficiency. The work provides a scalable, practical solution for robust PCC under diverse sensing conditions and domain shifts.

Abstract

Domain adaptive point cloud completion (DA PCC) aims to narrow the geometric and semantic discrepancies between the labeled source and unlabeled target domains. Existing methods either suffer from limited receptive fields or quadratic complexity due to using CNNs or vision Transformers. In this paper, we present the first work that studies the adaptability of State Space Models (SSMs) in DA PCC and find that directly applying SSMs to DA PCC will encounter several challenges: directly serializing 3D point clouds into 1D sequences often disrupts the spatial topology and local geometric features of the target domain. Besides, the overlook of designs in the learning domain-agnostic representations hinders the adaptation performance. To address these issues, we propose a novel framework, DAPointMamba for DA PCC, that exhibits strong adaptability across domains and has the advantages of global receptive fields and efficient linear complexity. It has three novel modules. In particular, Cross-Domain Patch-Level Scanning introduces patch-level geometric correspondences, enabling effective local alignment. Cross-Domain Spatial SSM Alignment further strengthens spatial consistency by modulating patch features based on cross-domain similarity, effectively mitigating fine-grained structural discrepancies. Cross-Domain Channel SSM Alignment actively addresses global semantic gaps by interleaving and aligning feature channels. Extensive experiments on both synthetic and real-world benchmarks demonstrate that our DAPointMamba outperforms state-of-the-art methods with less computational complexity and inference latency.

Figures (4)

• Figure 1: Left: Previous domain adaptive point cloud completion works suffer from limited receptive fields (a) or quadratic complexity (b) due to using CNNs or vision Transformers. Middle: In contrast, we propose a novel framework, DAPointMamba (c), that consists of Cross-Domain Patch-Level Scanning (CDPS), Cross-Domain Spatial SSM Alignment (CDSA), and Cross-Domain Channel SSM Alignment (CDCA). Our model exhibits strong adaptability across domains and has the advantages of global receptive fields and efficient linear complexity. Right: Visualization of the Table category from widely-used 3D-FUTURE benchmark, and our DAPointMamba demonstrates superior domain adaptability (lower Chamfer Distance is better) against state-of-the-art methods.
• Figure 2: The framework of DAPointMamba for cross-domain point cloud completion, including three key components: (a) Cross-Domain Patch-Level Scanning is designed to close domain shifts by creating spatial correspondence at the patch level. (b) Cross-Domain Spatial SSM Alignment is proposed to solve fine-grained spatial discrepancies across domains. (c) Cross-Domain Channel SSM Alignment is presented to address the semantic structure of feature channels through cross-domain channel mixing and similarity-based modulation.
• Figure 3: Visualization comparisons between ours and state-of-the-art PCC methods on 3D-FUTURE dataset.
• Figure 4: Visualization of the feature distribution of Cross-Domain Patch-Level Scanning (CDPS) compared with the global grouping scan strategy.