M2-SMap: Memory-Efficient Semantic Mapping with Hierarchical Multi-Model Representation

2026-08-07

AI Summary

Dense point cloud mapping remains challenging to deploy on resource-constrained robots, as existing approaches either lack representational capacity or suffer from overfitting and object sticking due to local geometric dominance in multi-primitive models. This work proposes M2-SMap, a framework that decomposes RGB-D point clouds hierarchically into bounded planes (for large planar regions), superquadrics (for semantic objects), and Gaussian mixture models (for complex residual structures). By integrating projection-guided semantic labeling, object-aware Gaussian merging, and multi-scale feature extraction, M2-SMap enables joint geometric and semantic modeling. The method runs in real time (≥29.37 Hz) on three RGB-D sequences, reducing the average number of primitives by 18.7% compared to the best baseline and eliminating inter-frame object sticking entirely.

Abstract

Dense point cloud maps, as a typically used mapping representation, are difficult to deploy on resource-constrained robots because their memory consumption grows rapidly with scene scale. Although compact single-model representations reduce memory cost, their fixed geometric expressiveness is insufficient for structurally diverse environments. Existing multi-model methods improve representational flexibility, yet their feature extraction and model selection are often dominated by local geometry, which can cause overfitting and adhesion between objects. To address these issues, this paper presents M2-SMap, a memory-efficient semantic mapping framework based on hierarchical multi-model representation. First, a hierarchical geometric decomposition partitions RGB-D point clouds into compact Gaussian components. Then, a projection-guided semantic annotation mechanism assigns instance identities to each component. Subsequently, these annotations are incorporated into an object-aware Gaussian fusion strategy. Furthermore, a multi-scale feature extraction strategy separates large planar regions, semantic objects, and complex residual structures, which are respectively represented by bounded planes, object-level superquadrics, and GMM primitives. Experiments on three RGB-D sequences show that M2-SMap runs in real time at no less than 29.37 Hz while achieving the lowest primitive count, with an average reduction of 18.7% over the best baseline. It also reduces the mean per-frame number of measured inter-object adhesion cases from 2.808 to 0, demonstrating efficient and semantically consistent scene representation.

Problem

• semantic mapping
• memory efficiency
• multi-model representation
• geometric expressiveness
• object adhesion

Innovation

• hierarchical multi-model representation
• memory-efficient semantic mapping
• object-aware Gaussian fusion
• multi-scale feature extraction
• semantic instance annotation