SpotOcc

To solve the decoder bottleneck in sparse 3D occupancy prediction, we propose SPOT-Occ. It uses a prototype-based sparse Transformer decoder, replacing costly dense attention with a "guided selection + focused aggregation" process. Queries focus on a small set of key "prototypes". The final model is significantly faster and more accurate.

Generate Convert Improve

Install / Use

npx skills add chensuzeyu/SpotOcc

Installs into whichever agent you are using.

About this skill

Quality Score

0/100

README

SPOT-Occ: Sparse Prototype-guided Transformer for Camera-based 3D Occupancy Prediction

Abstract

Achieving highly accurate and real-time 3D occupancy prediction from cameras is a critical requirement for the safe and practical deployment of autonomous vehicles. While this shift to sparse 3D representations solves the encoding bottleneck, it creates a new challenge for the decoder: how to efficiently aggregate information from a sparse, non-uniformly distributed set of voxel features without resorting to computationally prohibitive dense attention. In this paper, we propose a novel Prototype-based Sparse Transformer Decoder that replaces this costly interaction with an efficient, two-stage process of guided feature selection and focused aggregation. Our core idea is to make the decoder's attention prototype-guided. We achieve this through a sparse prototype selection mechanism, where each query adaptively identifies a compact set of the most salient voxel features, termed prototypes, for focused feature aggregation. To ensure this dynamic selection is stable and effective, we introduce a complementary denoising paradigm. This approach leverages ground-truth masks to provide explicit guidance, guaranteeing a consistent query-prototype association across decoder layers. Our model, dubbed SPOT-Occ, outperforms previous methods with a significant margin in speed while also improving accuracy.

model

spot-ca

Demo

demo

Benchmark Results

Occupancy Prediction on OpenOccupancy validation set: openocc-val

Semantic Scene Completion on SemanticKITTI validation set: kitti-val

Model Zoo

We provide the pretrained weights on SemanticKITTI and nuScenes datasets.

Getting Started

Citation

If you find this work useful, please consider citing:

@article{spotocc2026,
  title={SPOT-Occ: Sparse Prototype-guided Transformer for Camera-based 3D Occupancy Prediction},
  author={Chen, Suzeyu and Li, Leheng and Chen, Ying-Cong},
  journal={arXiv preprint arXiv:2602.04240},
  year={2026}
}

Acknowledgement

This project is developed based on the following open-sourced projects: BEVDet, BEVFormer, Mask2Former, OccFormer, OpenOccupancy, SparseOcc. Thanks for their excellent work.