---
license: mit
pipeline_tag: image-segmentation
---

# LaSSM: Efficient Semantic-Spatial Query Decoding via Local Aggregation and State Space Models for 3D Instance Segmentation

This repository contains the weights for **LaSSM**, a framework for efficient 3D scene instance segmentation from point clouds.

- **Paper:** [LaSSM: Efficient Semantic-Spatial Query Decoding via Local Aggregation and State Space Models for 3D Instance Segmentation](https://huggingface.co/papers/2602.11007)
- **Code:** [GitHub - RayYoh/LaSSM](https://github.com/RayYoh/LaSSM)

## Description
LaSSM (Local Aggregation and State Space Models) prioritizes simplicity and efficiency in 3D instance segmentation. It introduces a hierarchical semantic-spatial query initializer and a coordinate-guided state space model (SSM) decoder. This design restricts the model to focus on geometrically coherent regions and uses a spatial dual-path SSM block to capture underlying dependencies within the query set, reducing redundant computation.

LaSSM ranks first on the ScanNet++ V2 leaderboard, outperforming previous state-of-the-art methods with significantly fewer FLOPs.

## Trained Results
| Model | Benchmark | Num GPUs | mAP | AP50 | AP25 | Config | Tensorboard | Exp Record | Model |
| :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| LaSSM | ScanNet++ V2 Val | 4 | 29.1 | 43.5 | 51.6 | [Link](https://github.com/RayYoh/LaSSM/blob/main/configs/scannetpp/insseg-lassm-spunet-v2-3.py) | [Link](https://huggingface.co/RayYoh/LaSSM/tensorboard) | [Link](https://huggingface.co/RayYoh/LaSSM/raw/main/scannetpp-lassm-spunet-v2-3/train.log) | [Link](https://huggingface.co/RayYoh/LaSSM/blob/main/scannetpp-lassm-spunet-v2-3/model/model_best.pth) |
| LaSSM | ScanNet Val | 4 | 58.4 | 78.1 | 86.1 | [Link](https://github.com/RayYoh/LaSSM/blob/main/configs/scannet/insseg-lassm-spunet-v2-3.py) | - | [Link](https://huggingface.co/RayYoh/LaSSM/raw/main/scannet-lassm-spunet-v2-3/train.log) | [Link](https://huggingface.co/RayYoh/LaSSM/blob/main/scannet-lassm-spunet-v2-3/model/model_best.pth) |
| LaSSM | ScanNet200 Val | 4 | 29.3 | 39.2 | 44.5 | [Link](https://github.com/RayYoh/LaSSM/blob/main/configs/scannet200/insseg-lassm-minkunet-3.py) | - | [Link](https://huggingface.co/RayYoh/LaSSM/raw/main/scannet200-lassm-minkunet-3/train.log) | [Link](https://huggingface.co/RayYoh/LaSSM/blob/main/scannet200-lassm-minkunet-3/model/model_best.pth) |

## Citation
```bibtex
@article{yao2025lassm,
  title={LaSSM: Efficient Semantic-Spatial Query Decoding via Local Aggregation and State Space Models for 3D Instance Segmentation},
  author={Yao, Lei and Wang, Yi and Yawen, Cui and Liu, Moyun and Chau, Lap-Pui},
  journal={arXiv preprint arXiv:2602.11007},
  year={2025}
}
```