| title | SPARGen: Unifying Spatial Perception and Reasoning through Native Multimodal Generation |
|---|---|
| authors | Jinsheng Quan, Jianhua Li, Siyi Xie, Xuanke Shi, Kewang Deng, Zukai Chen, Feifei Shao, Lei Yang, Quan Wang, Yawei Luo |
| date | 2026-08-14 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2608.14138 |
| upvotes | 1 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Spatial perception and reasoning from visual observations require recovering geometric structure, establishing correspondences, and understanding spatial relations. Existing approaches typically address these capabilities separately using task-specific architectures or external geometric modules, limiting knowledge transfer among complementary representations of the same physical scene. We introduce SPARGen, a unified multimodal framework that casts 3D reconstruction, dense correspondence, and spatial reasoning as instruction-conditioned generation tasks. SPARGen serializes compact structured and linguistic outputs as token sequences while generating dense geometric fields in image-aligned forms, enabling spatial supervision to jointly shape shared representations within a native multimodal generative model. Experiments across benchmarks for 3D reconstruction, correspondence, and spatial reasoning show that SPARGen achieves competitive performance across heterogeneous spatial tasks within a single native multimodal generative framework. |
cite
@misc{quan2026spargen,
title={SPARGen: Unifying Spatial Perception and Reasoning through Native Multimodal Generation},
author={Jinsheng Quan and Jianhua Li and Siyi Xie and Xuanke Shi and Kewang Deng and Zukai Chen and Feifei Shao and Lei Yang and Quan Wang and Yawei Luo},
year={2026},
eprint={2608.14138},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2608.14138}
}