Paper
SoPE: Spherical Coordinate-Based Positional Embedding for Enhancing Spatial Perception of 3D LVLMs
Authors
Guanting Ye, Qiyan Zhao, Wenhao Yu, Liangyu Yuan, Mingkai Li, Xiaofeng Zhang, Jianmin Ji, Yanyong Zhang, Qing Jiang, Ka-Veng Yuen
Abstract
3D Large Vision-Language Models (3D LVLMs) built upon Large Language Models (LLMs) have achieved remarkable progress across various multimodal tasks. However, their inherited position-dependent modeling mechanism, Rotary Position Embedding (RoPE), remains suboptimal for 3D multimodal understanding. The vanilla RoPE formulation fails to preserve essential three-dimensional spatial structures when encoding 3D tokens, and its relative distance computation overlooks angular dependencies, hindering the model's ability to capture directional variations in visual representations. To overcome these limitations, we introduce Spherical Coordinate-based Positional Embedding (SoPE). Our method maps point-cloud token indices into a 3D spherical coordinate space, enabling unified modeling of spatial locations and directional angles. This formulation preserves the inherent geometric structure of point-cloud data, enhances spatial awareness, and yields more consistent and expressive geometric representations for multimodal learning. In addition, we introduce a multi-scale frequency mixing strategy to fuse feature information across different frequency domains. Experimental results on multiple 3D scene benchmarks validate the effectiveness of our approach, while real-world deployment experiments further demonstrate its strong generalization capability.
Metadata
Related papers
Vibe Coding XR: Accelerating AI + XR Prototyping with XR Blocks and Gemini
Ruofei Du, Benjamin Hersh, David Li, Nels Numan, Xun Qian, Yanhe Chen, Zhongy... • 2026-03-25
Comparing Developer and LLM Biases in Code Evaluation
Aditya Mittal, Ryan Shar, Zichu Wu, Shyam Agarwal, Tongshuang Wu, Chris Donah... • 2026-03-25
The Stochastic Gap: A Markovian Framework for Pre-Deployment Reliability and Oversight-Cost Auditing in Agentic Artificial Intelligence
Biplab Pal, Santanu Bhattacharya • 2026-03-25
Retrieval Improvements Do Not Guarantee Better Answers: A Study of RAG for AI Policy QA
Saahil Mathur, Ryan David Rittner, Vedant Ajit Thakur, Daniel Stuart Schiff, ... • 2026-03-25
MARCH: Multi-Agent Reinforced Self-Check for LLM Hallucination
Zhuo Li, Yupeng Zhang, Pengyu Cheng, Jiajun Song, Mengyu Zhou, Hao Li, Shujie... • 2026-03-25
Raw Data (Debug)
{
"raw_xml": "<entry>\n <id>http://arxiv.org/abs/2602.22716v1</id>\n <title>SoPE: Spherical Coordinate-Based Positional Embedding for Enhancing Spatial Perception of 3D LVLMs</title>\n <updated>2026-02-26T07:42:15Z</updated>\n <link href='https://arxiv.org/abs/2602.22716v1' rel='alternate' type='text/html'/>\n <link href='https://arxiv.org/pdf/2602.22716v1' rel='related' title='pdf' type='application/pdf'/>\n <summary>3D Large Vision-Language Models (3D LVLMs) built upon Large Language Models (LLMs) have achieved remarkable progress across various multimodal tasks. However, their inherited position-dependent modeling mechanism, Rotary Position Embedding (RoPE), remains suboptimal for 3D multimodal understanding. The vanilla RoPE formulation fails to preserve essential three-dimensional spatial structures when encoding 3D tokens, and its relative distance computation overlooks angular dependencies, hindering the model's ability to capture directional variations in visual representations. To overcome these limitations, we introduce Spherical Coordinate-based Positional Embedding (SoPE). Our method maps point-cloud token indices into a 3D spherical coordinate space, enabling unified modeling of spatial locations and directional angles. This formulation preserves the inherent geometric structure of point-cloud data, enhances spatial awareness, and yields more consistent and expressive geometric representations for multimodal learning. In addition, we introduce a multi-scale frequency mixing strategy to fuse feature information across different frequency domains. Experimental results on multiple 3D scene benchmarks validate the effectiveness of our approach, while real-world deployment experiments further demonstrate its strong generalization capability.</summary>\n <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n <published>2026-02-26T07:42:15Z</published>\n <arxiv:comment>CVPR 2026</arxiv:comment>\n <arxiv:primary_category term='cs.CV'/>\n <author>\n <name>Guanting Ye</name>\n </author>\n <author>\n <name>Qiyan Zhao</name>\n </author>\n <author>\n <name>Wenhao Yu</name>\n </author>\n <author>\n <name>Liangyu Yuan</name>\n </author>\n <author>\n <name>Mingkai Li</name>\n </author>\n <author>\n <name>Xiaofeng Zhang</name>\n </author>\n <author>\n <name>Jianmin Ji</name>\n </author>\n <author>\n <name>Yanyong Zhang</name>\n </author>\n <author>\n <name>Qing Jiang</name>\n </author>\n <author>\n <name>Ka-Veng Yuen</name>\n </author>\n </entry>"
}