# Point2Radio: A Foundation Model for Cross-Scene Radio Fields from Material-Aware Point Clouds

arXiv: https://arxiv.org/abs/2607.28994 · HTML: https://mlsanity.com/paper/2607.28994/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 32 references that are in the map.
- First posted on arXiv 31 Jul 2026.

## Builds on (most cited first, 21 of 21)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [SAM 3: Segment Anything with Concepts](https://mlsanity.com/paper/2511.16719/) (2025)
- [Infinigen Indoors: Photorealistic Indoor Scenes using Procedural Generation](https://mlsanity.com/paper/2406.11824/) (2024)
- [RadioUNet: Fast Radio Map Estimation with Convolutional Neural Networks](https://mlsanity.com/paper/1911.09002/) (2019)
- [A Tutorial on Environment-Aware Communications via Channel Knowledge Map for 6G](https://mlsanity.com/paper/2309.07460/) (2023)
- [NeRF2: Neural Radio-Frequency Radiance Fields](https://mlsanity.com/paper/2305.06118/) (2023)
- [WiFo: Wireless Foundation Model for Channel Prediction](https://mlsanity.com/paper/2412.08908/) (2024)
- [NeWRF: A Deep Learning Framework for Wireless Radiation Field Reconstruction and Channel Prediction](https://mlsanity.com/paper/2403.03241/) (2024)
- [RF-3DGS: Wireless Channel Modeling with Radio Radiance Field and 3D Gaussian Splatting](https://mlsanity.com/paper/2411.19420/) (2024)
- [PMNet: Robust Pathloss Map Prediction via Supervised Learning](https://mlsanity.com/paper/2211.10527/) (2022)
- [RadioDiff-3D: A 3D×3D Radio Map Dataset and Generative Diffusion Based Benchmark for 6G Environment-Aware Communication](https://mlsanity.com/paper/2507.12166/) (2025)
- [GSRF: Complex-Valued 3D Gaussian Splatting for Efficient Radio-Frequency Data Synthesis](https://mlsanity.com/paper/2502.01826/) (2025)
- [Radio Frequency Ray Tracing with Neural Object Representation](https://mlsanity.com/paper/2411.18635/) (2024)
- [OmniVLA: Physically-Grounded Multimodal VLA with Unified Multi-Sensor Perception for Robotic Manipulation](https://mlsanity.com/paper/2511.01210/) (2025)
- [RayProNet: A Neural Point Field Framework for Radio Propagation Modeling in 3D Environments](https://mlsanity.com/paper/2406.16907/) (2024)
- [Bridging Visual and Wireless Sensing via a Unified Radiation Field for 3D Radio Map Construction](https://mlsanity.com/paper/2601.19216/) (2026)
- [A Generalizable Light Transport 3D Embedding for Global Illumination](https://mlsanity.com/paper/2510.18189/) (2025)
- [Taming Vision Priors for Data Efficient mmWave Channel Modeling](https://mlsanity.com/paper/2603.13383/) (2026)
- [XFreq-GS: Cross-Frequency Wireless Radiation Field Reconstruction with 3D Gaussian Splatting](https://mlsanity.com/paper/2605.11432/) (2026)
- [GeoGS-CE: Learning Delay--Beam Channel Priors with 3D Gaussians for High-Mobility Scenarios](https://mlsanity.com/paper/2605.16094/) (2026)

## Built on by (most cited first, 1 of 1)

- [RFWeave: Learning Wireless Propagation Across Scenes and Transceiver Configurations](https://mlsanity.com/paper/2610.16656/) (2026)

## Abstract (arXiv)

High-fidelity radio fields are typically simulated for every scene--transmitter configuration or fitted separately to each scene, failing to exploit propagation structures shared across environments. We present Point2Radio, a foundation model that learns a transferable propagation prior from multiple environments. Given a material-aware point cloud and a transmitter (TX) setting, a common encoder produces a TX-conditioned scene representation that can be queried at arbitrary receiver (RX) locations. Task-specific query decoders map this representation to different radio quantities, e.g., three-dimensional (3D) path-gain (PG) fields and power angular spectra (PAS). At inference for a new scene, the model uses only a material-aware point cloud and transceiver queries, running in milliseconds on a single GPU without meshes or explicit path tracing. We evaluate PG prediction on a scene-disjoint split of a 337-scene corpus containing 86,272 TX-conditioned fields. Point2Radio achieves 0.871 dB mean absolute error (MAE), reducing error by 76.7% relative to a same-split UNet-style baseline. The same encoder also supports PAS prediction via a task-specific decoder. Experiments further show that light target-scene fine-tuning improves adaptation to a specific environment.