arxiv: 2604.02603 · v1 · submitted 2026-04-03 · 💻 cs.CV

Recognition: 2 theorem links

· Lean Theorem

Rascene: High-Fidelity 3D Scene Imaging with mmWave Communication Signals

Kunzhe Song , Geo Jie Zhou , Xiaoming Liu , Huacheng Zeng

Authors on Pith no claims yet

Pith reviewed 2026-05-13 20:06 UTC · model grok-4.3

classification 💻 cs.CV

keywords 3D scene reconstructionmmWave sensingISACmulti-frame fusionwireless imagingautonomous navigationradar imagingOFDM signals

0 comments

The pith

Rascene reconstructs high-precision 3D scenes from standard mmWave communication signals by fusing multiple frames with adaptive projection.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents Rascene as a way to turn ordinary millimeter-wave wireless signals into a source for detailed 3D environmental maps. Optical sensors like cameras and LiDAR often lose accuracy in fog, smoke, or darkness, while specialized radar hardware remains expensive and limited in spectrum access. Rascene addresses the sparsity and multipath confusion in single radio frames by combining data from several frames taken at different positions, using a spatially adaptive process that weights projections by confidence. A sympathetic reader would expect this to deliver reliable geometry recovery without extra calibration or custom equipment, opening 3D perception to everyday communication devices. Experimental validation shows the method achieves the precision needed for navigation tasks.

Core claim

Rascene is an integrated sensing and communication framework that uses ubiquitous mmWave OFDM signals for 3D scene imaging. Individual radio frames are sparse and multipath-ambiguous, so the system applies multi-frame, spatially adaptive fusion with confidence-weighted forward projection to recover geometric consensus across arbitrary poses and produce high-fidelity 3D reconstructions.

What carries the argument

multi-frame spatially adaptive fusion with confidence-weighted forward projection, which aligns signals from different transmitter-receiver poses and combines them to extract consistent 3D geometry.

If this is right

Enables 3D perception for autonomous driving and robot navigation in smoke, fog, and non-ideal lighting.
Delivers low-cost 3D imaging by reusing existing mmWave communication hardware instead of dedicated radar.
Supports scalable deployment without requiring licensed spectrum or bespoke sensing equipment.
Allows reconstruction from signals collected at arbitrary poses without extra calibration steps.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same fusion approach could be applied to ambient 5G signals for city-scale passive mapping without active transmissions.
Accuracy might improve further if the method is combined with occasional optical measurements in a hybrid sensor suite.
Real-time versions would require optimizing the fusion pipeline for lower latency on embedded platforms.
Performance in highly dynamic scenes with moving objects remains an open extension beyond the static-scene tests.

Load-bearing premise

Multi-frame fusion of sparse, multipath-ambiguous radio frames can reliably recover accurate 3D geometry without specialized hardware or additional calibration.

What would settle it

Controlled experiments in which ground-truth 3D geometry is known show that Rascene point clouds deviate by more than a few centimeters from LiDAR references under realistic multipath conditions.

Figures

Figures reproduced from arXiv: 2604.02603 by Geo Jie Zhou, Huacheng Zeng, Kunzhe Song, Xiaoming Liu.

**Figure 1.** Figure 1: High-fidelity 3D imaging generated by Rascene from mmWave communication signals. We show that OFDM communication [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗

**Figure 2.** Figure 2: Illustration of monostatic sensing in a mmWave commu [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Figure 3: Illustration of angular estimation on a mmWave device. [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

**Figure 5.** Figure 5: Overview of the multi-frame 3D RF imaging network. Given multiple radio frames and poses, a shared encoder predicts per [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗

**Figure 6.** Figure 6: Rascene data collection platform. 5. Implementation Hardware Setup. We built a prototype of Rascene as shown in [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗

**Figure 7.** Figure 7: Qualitative results of our Rascene system. For each row, we show the ground truth depth map and voxel grid derived from [PITH_FULL_IMAGE:figures/full_fig_p006_7.png] view at source ↗

**Figure 8.** Figure 8: Qualitative comparison of single-frame and 5-frame predictions. Multi-frame fusion reduces missed detections and suppresses [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗

**Figure 9.** Figure 9: Cumulative distribution functions illustrating absolute [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗

**Figure 12.** Figure 12: Representative examples of different sensors’ occlusion [PITH_FULL_IMAGE:figures/full_fig_p008_12.png] view at source ↗

**Figure 11.** Figure 11: Depth error as a function of range for single-frame and [PITH_FULL_IMAGE:figures/full_fig_p008_11.png] view at source ↗

**Figure 14.** Figure 14: Illustration of video streaming data communication [PITH_FULL_IMAGE:figures/full_fig_p011_14.png] view at source ↗

**Figure 13.** Figure 13: Our prototyped monostatic ISAC device. 8. Monostatic ISAC Hardware We built a monostatic ISAC prototype using commercial off-the-shelf (COTS) components, enabling joint communication and sensing within a compact device [PITH_FULL_IMAGE:figures/full_fig_p011_13.png] view at source ↗

**Figure 15.** Figure 15: Sample trajectory segments (shown in red) from different scenes, visualized on the ground truth LiDAR point clouds. [PITH_FULL_IMAGE:figures/full_fig_p012_15.png] view at source ↗

**Figure 16.** Figure 16: Example snapshots from the 20 distinct indoor envi [PITH_FULL_IMAGE:figures/full_fig_p012_16.png] view at source ↗

read the original abstract

Robust 3D environmental perception is critical for applications such as autonomous driving and robot navigation. However, optical sensors such as cameras and LiDAR often fail under adverse conditions, including smoke, fog, and non-ideal lighting. Although specialized radar systems can operate in these environments, their reliance on bespoke hardware and licensed spectrum limits scalability and cost-effectiveness. This paper introduces Rascene, an integrated sensing and communication (ISAC) framework that leverages ubiquitous mmWave OFDM communication signals for 3D scene imaging. To overcome the sparse and multipath-ambiguous nature of individual radio frames, Rascene performs multi-frame, spatially adaptive fusion with confidence-weighted forward projection, enabling the recovery of geometric consensus across arbitrary poses. Experimental results demonstrate that our method reconstructs 3D scenes with high precision, offering a new pathway toward low-cost, scalable, and robust 3D perception.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Rascene sketches an ISAC route to 3D imaging from ordinary mmWave OFDM signals via multi-frame fusion, but the precision claims rest on thin validation.

read the letter

The paper's main move is to repurpose standard mmWave communication waveforms for 3D scene recovery instead of relying on dedicated radar hardware. It does this by stacking multiple frames and applying spatially adaptive fusion with a confidence-weighted forward projection step that tries to find geometric agreement across different viewpoints. That combination is the concrete new piece; prior ISAC work has looked at sensing from comm signals, but the specific multi-frame handling of pose variation and ambiguity looks like a fresh application angle here. The motivation is solid too: optical sensors drop out in fog or smoke, and licensed radar is expensive, so using existing infrastructure is a practical direction if it works at scale. The abstract states that experiments show high-precision reconstruction, which is the claim worth checking. The soft spot is exactly where the stress-test note flags it. The method needs to resolve multipath and sparsity without extra calibration, yet the description gives no equations for how the weights are derived, how the projection copes with arbitrary poses, or any numerical error metrics, baselines, or ablation results. Without those, the central assertion that geometric consensus emerges reliably stays hard to evaluate. The paper is aimed at researchers in integrated sensing and communication or perception for robotics and vehicles. A reader already working on mmWave ISAC could pick up the fusion idea and test it, but anyone expecting ready-to-use high-fidelity results will need the full experimental section to decide. I would send it to peer review so the methods and data can be examined properly; the idea has enough grounding to merit that step even if revisions are needed on the validation side.

Referee Report

3 major / 1 minor

Summary. The paper introduces Rascene, an ISAC framework that uses ubiquitous mmWave OFDM communication signals for 3D scene imaging. To address the sparse and multipath-ambiguous nature of individual radio frames, it performs multi-frame, spatially adaptive fusion with confidence-weighted forward projection to recover geometric consensus across arbitrary poses, claiming high-precision 3D reconstruction without specialized hardware or calibration.

Significance. If the central claims are substantiated with quantitative evidence, the work would offer a scalable, low-cost pathway for robust 3D perception in adverse conditions by repurposing existing communication infrastructure, potentially impacting autonomous driving and robotics where optical sensors fail.

major comments (3)

[Abstract] Abstract: the claim that the method 'reconstructs 3D scenes with high precision' is unsupported by any quantitative metrics, error analysis, or validation details, which is load-bearing for the central performance assertion.
[Method] Method (multi-frame fusion description): no explicit formulation is given for how confidence weights are computed or how forward projection resolves pose variation and multipath ambiguities, leaving the key disambiguation step unsubstantiated.
[Experimental results] Experimental results: the abstract states 'experimental results demonstrate' high precision, yet the absence of reported error metrics, baselines, or ablation studies on multipath handling prevents assessment of whether the fusion actually recovers accurate geometry without calibration.

minor comments (1)

[Abstract] Abstract: expand the acronym ISAC on first use for clarity.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We sincerely thank the referee for the thorough and constructive review. We address each major comment below and will revise the manuscript to incorporate the requested clarifications and quantitative support.

read point-by-point responses

Referee: [Abstract] Abstract: the claim that the method 'reconstructs 3D scenes with high precision' is unsupported by any quantitative metrics, error analysis, or validation details, which is load-bearing for the central performance assertion.

Authors: We agree that the abstract claim requires explicit quantitative backing. In the revised manuscript we will update the abstract to reference concrete metrics (e.g., mean point-to-point RMSE and standard deviation relative to ground-truth LiDAR) that appear in the experimental section, and we will add a short validation summary to make the performance assertion self-contained. revision: yes
Referee: [Method] Method (multi-frame fusion description): no explicit formulation is given for how confidence weights are computed or how forward projection resolves pose variation and multipath ambiguities, leaving the key disambiguation step unsubstantiated.

Authors: We acknowledge the need for explicit mathematics. We will insert the missing equations: confidence weights will be defined as a product of per-voxel SNR and cross-frame geometric consistency; forward projection will be formulated as a pose-transformed accumulation followed by a consensus filter that discards inconsistent multipath returns. These additions will directly substantiate the disambiguation mechanism. revision: yes
Referee: [Experimental results] Experimental results: the abstract states 'experimental results demonstrate' high precision, yet the absence of reported error metrics, baselines, or ablation studies on multipath handling prevents assessment of whether the fusion actually recovers accurate geometry without calibration.

Authors: We will expand the experimental section to include quantitative error metrics (RMSE, precision-recall curves), comparisons against single-frame and conventional radar baselines, and dedicated ablation studies isolating the multipath-handling and multi-frame fusion components. These additions will allow direct assessment of geometric accuracy without external calibration. revision: yes

Circularity Check

0 steps flagged

No circularity detected in derivation chain

full rationale

The abstract presents Rascene as a novel ISAC framework that applies multi-frame spatially adaptive fusion with confidence-weighted forward projection to mmWave OFDM signals. No equations, parameter fits, or self-citations are shown that would make any claimed prediction or geometric consensus equivalent to the inputs by construction. The method description remains independent of its outputs, with no self-definitional loops, fitted inputs renamed as predictions, or load-bearing self-citations. The derivation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Review based solely on abstract; no explicit free parameters, axioms, or invented entities are identifiable from the provided text.

pith-pipeline@v0.9.0 · 5459 in / 1059 out tokens · 33383 ms · 2026-05-13T20:06:18.120914+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Zr(xr) = sum wr←i(xi,xr) Fi(xi) / (sum wr←i + ε) with wr←i = K_σ(˜xr←i,xr) · [α(Ci(xi))]^η
IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

multi-frame, spatially adaptive fusion with confidence-weighted forward projection

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

51 extracted references · 51 canonical work pages

[1]

Enabling{high-quality}untethered virtual reality

Omid Abari, Dinesh Bharadia, Austin Duffield, and Dina Katabi. Enabling{high-quality}untethered virtual reality. In14th USENIX Symposium on Networked Systems Design and Implementation (NSDI 17), pages 531–544, 2017. 1, 2

work page 2017
[2]

3d tracking via body radio reflections

Fadel Adib, Zach Kabelac, Dina Katabi, and Robert C Miller. 3d tracking via body radio reflections. In11th USENIX Sym- posium on Networked Systems Design and Implementation (NSDI 14), pages 317–329, 2014

work page 2014
[3]

Capturing the human figure through a wall

Fadel Adib, Chen-Yu Hsu, Hongzi Mao, Dina Katabi, and Fr´edo Durand. Capturing the human figure through a wall. ACM Transactions on Graphics (TOG), 34(6):1–13, 2015. 1

work page 2015
[4]

In12th USENIX Symposium on Networked Systems Design and Im- plementation (NSDI 15), pages 279–292, 2015

Fadel Adib, Zachary Kabelac, and Dina Katabi.{Multi- Person}localization via{RF}body reflections. In12th USENIX Symposium on Networked Systems Design and Im- plementation (NSDI 15), pages 279–292, 2015. 2

work page 2015
[5]

Orb-slam3: An accurate open-source library for visual, visual–inertial, and multimap slam.IEEE transactions on robotics, 37(6):1874– 1890, 2021

Carlos Campos, Richard Elvira, Juan J G ´omez Rodr´ıguez, Jos´e MM Montiel, and Juan D Tard ´os. Orb-slam3: An accurate open-source library for visual, visual–inertial, and multimap slam.IEEE transactions on robotics, 37(6):1874– 1890, 2021. 2

work page 2021
[6]

Suma++: Efficient lidar-based semantic slam

Xieyuanli Chen, Andres Milioto, Emanuele Palazzolo, Philippe Giguere, Jens Behley, and Cyrill Stachniss. Suma++: Efficient lidar-based semantic slam. In2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 4530–4537. IEEE, 2019. 1, 2

work page 2019
[7]

Stachniss

Xieyuanli Chen, Thomas L ¨abe, Andres Milioto, Timo R¨ohling, Olga Vysotska, Alexandre Haag, Jens Behley, and C. Stachniss. Overlapnet: Loop closing for lidar-based slam. ArXiv, abs/2105.11344, 2020. 1

work page arXiv 2020
[8]

Learning implicit fields for generative shape modeling

Zhiqin Chen and Hao Zhang. Learning implicit fields for generative shape modeling. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5939–5948, 2019. 2

work page 2019
[9]

Integrated sensing and communications (isac) for ve- hicular communication networks (vcn).IEEE Internet of Things Journal, 9(23):23441–23451, 2022

Xiang Cheng, Dongliang Duan, Shijian Gao, and Liuqing Yang. Integrated sensing and communications (isac) for ve- hicular communication networks (vcn).IEEE Internet of Things Journal, 9(23):23441–23451, 2022. 2

work page 2022
[10]

A novel radar point cloud generation method for robot envi- ronment perception.IEEE Transactions on Robotics, 38(6): 3754–3773, 2022

Yuwei Cheng, Jingran Su, Mengxin Jiang, and Yimin Liu. A novel radar point cloud generation method for robot envi- ronment perception.IEEE Transactions on Robotics, 38(6): 3754–3773, 2022. 1, 2

work page 2022
[11]

Implicit functions in feature space for 3d shape reconstruc- tion and completion

Julian Chibane, Thiemo Alldieck, and Gerard Pons-Moll. Implicit functions in feature space for 3d shape reconstruc- tion and completion. InProceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, pages 6970–6981, 2020. 2

work page 2020
[12]

Neural unsigned distance fields for implicit function learning.Advances in Neural Information Processing Systems, 33:21638–21652,

Julian Chibane, Gerard Pons-Moll, et al. Neural unsigned distance fields for implicit function learning.Advances in Neural Information Processing Systems, 33:21638–21652,

work page
[13]

Isac overview.Hyperfine Interactions, 225(1):1–8, 2014

J Dilling, R Kr ¨ucken, and G Ball. Isac overview.Hyperfine Interactions, 225(1):1–8, 2014. 2

work page 2014
[14]

Sensing as a service in 6g percep- tive networks: A unified framework for isac resource allo- cation.IEEE Transactions on Wireless Communications, 22 (5):3522–3536, 2022

Fuwang Dong, Fan Liu, Yuanhao Cui, Wei Wang, Kaifeng Han, and Zhiqin Wang. Sensing as a service in 6g percep- tive networks: A unified framework for isac resource allo- cation.IEEE Transactions on Wireless Communications, 22 (5):3522–3536, 2022. 2

work page 2022
[15]

Efficient continuous- time slam for 3d lidar-based online mapping

David Droeschel and Sven Behnke. Efficient continuous- time slam for 3d lidar-based online mapping. In2018 IEEE International Conference on Robotics and Automation (ICRA), pages 5000–5007. IEEE, 2018. 2

work page 2018
[16]

Lsd- slam: Large-scale direct monocular slam

Jakob Engel, Thomas Sch ¨ops, and Daniel Cremers. Lsd- slam: Large-scale direct monocular slam. InEuropean con- ference on computer vision, pages 834–849. Springer, 2014. 2

work page 2014
[17]

Direct sparse odometry.IEEE transactions on pattern analysis and machine intelligence, 40(3):611–625, 2017

Jakob Engel, Vladlen Koltun, and Daniel Cremers. Direct sparse odometry.IEEE transactions on pattern analysis and machine intelligence, 40(3):611–625, 2017. 1, 2

work page 2017
[18]

Learning shape templates with structured implicit functions

Kyle Genova, Forrester Cole, Daniel Vlasic, Aaron Sarna, William T Freeman, and Thomas Funkhouser. Learning shape templates with structured implicit functions. InPro- ceedings of the IEEE/CVF international conference on com- puter vision, pages 7154–7164, 2019. 2

work page 2019
[19]

Through fog high-resolution imag- ing using millimeter wave radar

Junfeng Guan, Sohrab Madani, Suraj Jog, Saurabh Gupta, and Haitham Hassanieh. Through fog high-resolution imag- ing using millimeter wave radar. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11464–11473, 2020. 1

work page 2020
[20]

Eldar, and Xiaohu You

Zhenyao He, Wei Xu, Hong Shen, Derrick Wing Kwan Ng, Yonina C. Eldar, and Xiaohu You. Full-duplex communi- cation for isac: Joint beamforming and power optimization. IEEE Journal on Selected Areas in Communications, 41(9): 2920–2936, 2023. 2

work page 2023
[21]

Real-time loop closure in 2d lidar slam

Wolfgang Hess, Damon Kohler, Holger Rapp, and Daniel Andor. Real-time loop closure in 2d lidar slam. In2016 IEEE international conference on robotics and automation (ICRA), pages 1271–1278. IEEE, 2016. 1, 2

work page 2016
[22]

Point-to-voxel knowledge distillation for lidar se- mantic segmentation

Yuenan Hou, Xinge Zhu, Yuexin Ma, Chen Change Loy, and Yikang Li. Point-to-voxel knowledge distillation for lidar se- mantic segmentation. InProceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, pages 8479–8488, 2022. 2

work page 2022
[23]

Towards foundational models for single-chip radar

Tianshu Huang, Akarsh Prabhakara, Chuhan Chen, Jay Karhade, Deva Ramanan, Matthew O’toole, and Anthony Rowe. Towards foundational models for single-chip radar. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 24655–24665, 2025. 1

work page 2025
[24]

3d gaussian splatting for real-time radiance field rendering.ACM Trans

Bernhard Kerbl, Georgios Kopanas, Thomas Leimk ¨uhler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering.ACM Trans. Graph., 42(4):139–1,

work page
[25]

Fundamental trade-offs in monostatic isac: A holistic investigation to- wards 6g.IEEE Transactions on Wireless Communications,

Musa Furkan Keskin, Mohammad Mahdi Mojahedian, Je- sus O Lacruz, Carina Marcus, Olof Eriksson, Andrea Gior- getti, Joerg Widmer, and Henk Wymeersch. Fundamental trade-offs in monostatic isac: A holistic investigation to- wards 6g.IEEE Transactions on Wireless Communications,

work page
[26]

Spotfi: Decimeter level localization using wifi

Manikanta Kotaru, Kiran Joshi, Dinesh Bharadia, and Sachin Katti. Spotfi: Decimeter level localization using wifi. In Proceedings of the 2015 ACM conference on special interest group on data communication, pages 269–282, 2015. 2 9

work page 2015
[27]

Enabling visual recognition at radio frequency

Haowen Lai, Gaoxiang Luo, Yifei Liu, and Mingmin Zhao. Enabling visual recognition at radio frequency. InProceed- ings of the 30th Annual International Conference on Mobile Computing and Networking, pages 388–403, 2024. 1, 2, 7

work page 2024
[28]

Rf-based 3d slam rivaling vision approaches

Haowen Lai, Zhiwei Zheng, and Mingmin Zhao. Rf-based 3d slam rivaling vision approaches. InProceedings of the 31th Annual International Conference on Mobile Computing and Networking (MobiCom), pages 170–185, 2025. 7

work page 2025
[29]

Tay- lor, Mathias Unberath, Ming-Yu Liu, and Chen-Hsuan Lin

Zhaoshuo Li, Thomas M ¨uller, Alex Evans, Russell H. Tay- lor, Mathias Unberath, Ming-Yu Liu, and Chen-Hsuan Lin. Neuralangelo: High-fidelity neural surface reconstruction. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 8456–8465,

work page
[30]

Megasam: Accurate, fast and robust structure and motion from casual dynamic videos

Zhengqi Li, Richard Tucker, Forrester Cole, Qianqian Wang, Linyi Jin, Vickie Ye, Angjoo Kanazawa, Aleksander Holyn- ski, and Noah Snavely. Megasam: Accurate, fast and robust structure and motion from casual dynamic videos. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10486–10496, 2025. 1

work page 2025
[31]

Toward integrated sensing and com- munications in ieee 802.11 bf wi-fi networks.IEEE Commu- nications Magazine, 61(7):128–133, 2023

Francesca Meneghello, Cheng Chen, Carlos Cordeiro, and Francesco Restuccia. Toward integrated sensing and com- munications in ieee 802.11 bf wi-fi networks.IEEE Commu- nications Magazine, 61(7):128–133, 2023. 2

work page 2023
[32]

Occupancy networks: Learning 3d reconstruction in function space

Lars Mescheder, Michael Oechsle, Michael Niemeyer, Se- bastian Nowozin, and Andreas Geiger. Occupancy networks: Learning 3d reconstruction in function space. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4460–4470, 2019. 2

work page 2019
[33]

Nerf: Representing scenes as neural radiance fields for view syn- thesis.Communications of the ACM, 65(1):99–106, 2021

Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view syn- thesis.Communications of the ACM, 65(1):99–106, 2021. 1, 2

work page 2021
[34]

Dtam: Dense tracking and mapping in real-time

Richard A Newcombe, Steven J Lovegrove, and Andrew J Davison. Dtam: Dense tracking and mapping in real-time. In2011 international conference on computer vision, pages 2320–2327. IEEE, 2011. 2

work page 2011
[35]

Mulls: Versatile lidar slam via multi-metric lin- ear least square

Yue Pan, Pengchuan Xiao, Yujie He, Zhenlei Shao, and Zesong Li. Mulls: Versatile lidar slam via multi-metric lin- ear least square. In2021 IEEE International Conference on Robotics and Automation (ICRA), pages 11633–11640. IEEE, 2021. 2

work page 2021
[36]

Deepsdf: Learning con- tinuous signed distance functions for shape representation

Jeong Joon Park, Peter Florence, Julian Straub, Richard Newcombe, and Steven Lovegrove. Deepsdf: Learning con- tinuous signed distance functions for shape representation. InProceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 165–174, 2019. 2

work page 2019
[37]

High resolution point clouds from mmwave radar

Akarsh Prabhakara, Tao Jin, Arnav Das, Gantavya Bhatt, Lilly Kumari, Elahe Soltanaghai, Jeff Bilmes, Swarun Ku- mar, and Anthony Rowe. High resolution point clouds from mmwave radar. In2023 IEEE International Conference on Robotics and Automation (ICRA), pages 4135–4142, 2023. 1, 2

work page 2023
[38]

Semantic scene completion using local deep implicit functions on lidar data.IEEE transactions on pattern analysis and machine intelligence, 44(10):7205– 7218, 2021

Christoph B Rist, David Emmerichs, Markus Enzweiler, and Dariu M Gavrila. Semantic scene completion using local deep implicit functions on lidar data.IEEE transactions on pattern analysis and machine intelligence, 44(10):7205– 7218, 2021. 2

work page 2021
[39]

Multipath triangulation: Decimeter-level wifi localization and orientation with a single unaided receiver

Elahe Soltanaghaei, Avinash Kalyanaraman, and Kamin Whitehouse. Multipath triangulation: Decimeter-level wifi localization and orientation with a single unaided receiver. In Proceedings of the 16th annual international conference on mobile systems, applications, and services, pages 376–388,

work page
[40]

Siwis: Fine-grained human detection using single wifi device

Kunzhe Song, Qijun Wang, Shichen Zhang, and Huacheng Zeng. Siwis: Fine-grained human detection using single wifi device. InProceedings of the 30th Annual International Con- ference on Mobile Computing and Networking, pages 1439– 1454, 2024. 2

work page 2024
[41]

Spectrum shortage for radio sensing? leveraging ambient 5g signals for human activity detection.arXiv preprint arXiv:2603.03579, 2026

Kunzhe Song, Maxime Zingraff, and Huacheng Zeng. Spectrum shortage for radio sensing? leveraging ambient 5g signals for human activity detection.arXiv preprint arXiv:2603.03579, 2026. 2

work page arXiv 2026
[42]

mmtrack: Passive multi-person localization using commod- ity millimeter wave radio

Chenshu Wu, Feng Zhang, Beibei Wang, and KJ Ray Liu. mmtrack: Passive multi-person localization using commod- ity millimeter wave radio. InIEEE INFOCOM 2020-IEEE Conference on Computer Communications, pages 2400–

work page 2020
[43]

4d gaussian splatting for real-time dynamic scene render- ing

Guanjun Wu, Taoran Yi, Jiemin Fang, Lingxi Xie, Xiaopeng Zhang, Wei Wei, Wenyu Liu, Qi Tian, and Xinggang Wang. 4d gaussian splatting for real-time dynamic scene render- ing. InProceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 20310– 20320, 2024. 2

work page 2024
[44]

Sensing in bistatic isac systems with clock asynchronism: A signal processing perspective.IEEE Signal Processing Mag- azine, 41(5):31–43, 2024

Kai Wu, Jacopo Pegoraro, Francesca Meneghello, J Andrew Zhang, Jesus O Lacruz, Joerg Widmer, Francesco Restuc- cia, Michele Rossi, Xiaojing Huang, Daqing Zhang, et al. Sensing in bistatic isac systems with clock asynchronism: A signal processing perspective.IEEE Signal Processing Mag- azine, 41(5):31–43, 2024. 2

work page 2024
[45]

Zhaoyang Xia, You-Chen Liu, Xin Li, Xinge Zhu, Yuexin Ma, Yikang Li, Yuenan Hou, and Y . Qiao. Scpnet: Semantic scene completion on point cloud.2023 IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR), pages 17642–17651, 2023. 2

work page 2023
[46]

Person-in-wifi 3d: End-to-end multi- person 3d pose estimation with wi-fi

Kangwei Yan, Fei Wang, Bo Qian, Han Ding, Jinsong Han, and Xing Wei. Person-in-wifi 3d: End-to-end multi- person 3d pose estimation with wi-fi. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 969–978, 2024. 2

work page 2024
[47]

Towards dense and accurate radar perception via efficient cross-modal diffusion model.IEEE Robotics and Automation Letters, 2024

Ruibin Zhang, Donglai Xue, Yuhan Wang, Ruixu Geng, and Fei Gao. Towards dense and accurate radar perception via efficient cross-modal diffusion model.IEEE Robotics and Automation Letters, 2024. 1, 2

work page 2024
[48]

Emotion recognition using wireless signals

Mingmin Zhao, Fadel Adib, and Dina Katabi. Emotion recognition using wireless signals. InProceedings of the 22nd annual international conference on mobile computing and networking, pages 95–108, 2016. 2

work page 2016
[49]

Cylindrical and asymmetrical 3d convolution networks for lidar segmenta- tion.2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 9934–9943, 2020

Xinge Zhu, Hui Zhou, Tai Wang, Fangzhou Hong, Yuexin Ma, Wei Li, Hongsheng Li, and Dahua Lin. Cylindrical and asymmetrical 3d convolution networks for lidar segmenta- tion.2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 9934–9943, 2020. 2 10 Rascene: High-Fidelity 3D Scene Imaging with mmWave Communication Signals Supplem...

work page 2021
[50]

Monostatic ISAC Hardware We built a monostatic ISAC prototype using commercial off-the-shelf (COTS) components, enabling joint commu- nication and sensing within a compact device. Fig. 13 shows our monostatic ISAC prototype, with its parameters summarized in Tab. 7. The system consists of two primary COTS modules: (i) an AMD/Xilinx RFSoC 4x2 FPGA board, a...

work page
[51]

Data Collection Platform.To collect paired RF-LiDAR data, we mounted our custom-designed ISAC device, an Ouster OS0-128 Li- DAR, and a TDK ICM-20948 IMU on a movable cart. The final dataset contains synchronized RF-LiDAR frame pairs collected from 20 indoor environments spanning di- verse layouts, clutter levels, and construction materials such as drywall...

work page