arxiv: 2604.16201 · v1 · submitted 2026-04-17 · 💻 cs.RO · cs.CV

Recognition: unknown

DENALI: A Dataset Enabling Non-Line-of-Sight Spatial Reasoning with Low-Cost LiDARs

Nikhil Behari , Diego Rivero , Luke Apostolides , Suman Ghosh , Paul Pu Liang , Ramesh Raskar

Authors on Pith no claims yet

Pith reviewed 2026-05-10 08:17 UTC · model grok-4.3

classification 💻 cs.RO cs.CV

keywords non-line-of-sightLiDARdatasethidden objectsdata-driven inferencetime-resolved histogramsNLOS perceptionconsumer sensors

0 comments

The pith

Low-cost LiDARs enable data-driven perception of hidden objects via their raw time-resolved histograms.

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

The paper presents DENALI as the first large-scale real-world collection of space-time histograms from consumer LiDARs in scenes containing hidden objects. It establishes that models can learn to infer non-line-of-sight information from these histograms despite hardware constraints that prevent traditional reconstruction methods. This approach matters because it opens a path for everyday devices like phones and robots to detect objects around corners or behind obstacles using existing sensors. The work also highlights practical limits from scene variations and differences between simulation and real captures that future efforts must address.

Core claim

We present DENALI, the first large-scale real-world dataset of space-time histograms from low-cost LiDARs capturing hidden objects. Using our dataset, we show that consumer LiDARs can enable accurate, data-driven NLOS perception. We further identify key scene and modeling factors that limit performance, as well as simulation-fidelity gaps that hinder current sim-to-real transfer.

What carries the argument

DENALI dataset of space-time histograms from low-cost LiDARs encoding multi-bounce light returns for hidden objects.

If this is right

Data-driven models trained on the histograms achieve accurate non-line-of-sight perception with consumer LiDARs.
Scene factors including object shape, position, lighting conditions, and resolution affect model performance.
Simulation-to-real gaps limit transfer, motivating more real-world data collection for better models.
Scalable non-line-of-sight vision systems become feasible for mobile devices and robots.

Where Pith is reading between the lines

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

Robots and mobile phones could gain the ability to sense hidden objects without extra hardware.
Algorithms might combine these NLOS cues with direct depth measurements for improved scene understanding.
Collecting similar datasets for dynamic or outdoor environments could extend the approach to new applications.

Load-bearing premise

Data-driven models trained on the captured histograms can generalize to achieve accurate NLOS perception across unseen scenes despite the severe hardware limitations of consumer LiDARs.

What would settle it

Testing a trained model on a collection of new hidden-object scenes not included in the 72,000 training examples shows low accuracy in predicting object presence or location.

Figures

Figures reproduced from arXiv: 2604.16201 by Diego Rivero, Luke Apostolides, Nikhil Behari, Paul Pu Liang, Ramesh Raskar, Suman Ghosh.

**Figure 1.** Figure 1: Enabling data-driven non-line-of-sight (NLOS) spatial reasoning with low-cost LiDARs. DENALI is a large-scale dataset of 72,000 full time-resolved histograms captured with low-cost LiDARs in scenes designed to elicit multi-bounce returns from hidden objects. The dataset spans 60 object shapes, 100 positions, two lighting conditions, and two LiDAR resolutions. Using these diverse captures, we demonstrate th… view at source ↗

**Figure 2.** Figure 2 [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Figure 3: Capture setup. Our capture system is designed to record a large-scale dataset of non-line-of-sight, three-bounce light signals from a mobile flash LiDAR. The setup includes a low-cost single-photon LiDAR co-located with an Intel RealSense RGB-D camera, both directed toward a flat relay wall. A hidden object is mounted on a motorized gantry positioned outside their direct line of sight to ensure only indire… view at source ↗

**Figure 4.** Figure 4: Captured dataset objects. We 3D print 30 objects (10 letters, 10 numbers, 10 shapes) at two scales (4in. and 8in.). Retroreflective tape is applied to improve three-bounce photon returns in the measured LiDAR histograms. Each object is printed from a known CAD model; as such, accurate ground-truth meshes are available for downstream analysis and simulation. 3.3. Digital Twin Capture Our dataset includes 6… view at source ↗

**Figure 5.** Figure 5: Summary of dataset captures (left) and sample three-bounce signals from low-cost LiDAR (right). Example histograms from the central pixel of 3 × 3 and 8 × 8 configurations illustrate variations in three-bounce signal: intensity differences by object size, temporal shifts by position, and changes in photon counts and noise across lighting conditions and LiDAR resolution. 4. Analysis We have three primary an… view at source ↗

**Figure 6.** Figure 6: Spatial mapping of NLOS localization accuracy. We plot RMSE (m) over true gantry positions for a single trained model (1D CNN) broken down by size/lighting. Accuracy generally improves for larger (8in.) objects nearer to relay wall. However, different lighting induces distinct spatial error patterns, suggesting poor separation of object, geometry, and lighting [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗

**Figure 7.** Figure 7: NLOS classification accuracy across object types. We train a 1D CNN independently for each object size and evaluate classification performance across shapes; larger 8-inch objects yield consistently higher accuracy. Convolutional architectures (e.g., 1D and 3D CNNs) consistently achieve the best performance across tasks (Tab. 2), indicating that an inductive bias toward local temporal structure is well s… view at source ↗

**Figure 9.** Figure 9: We find that DENALI provides a useful benchmark for evaluating simulation fidelity: it enables us to quantify how much localization error decreases for each simulation [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗

read the original abstract

Consumer LiDARs in mobile devices and robots typically output a single depth value per pixel. Yet internally, they record full time-resolved histograms containing direct and multi-bounce light returns; these multi-bounce returns encode rich non-line-of-sight (NLOS) cues that can enable perception of hidden objects in a scene. However, severe hardware limitations of consumer LiDARs make NLOS reconstruction with conventional methods difficult. In this work, we motivate a complementary direction: enabling NLOS perception with low-cost LiDARs through data-driven inference. We present DENALI, the first large-scale real-world dataset of space-time histograms from low-cost LiDARs capturing hidden objects. We capture time-resolved LiDAR histograms for 72,000 hidden-object scenes across diverse object shapes, positions, lighting conditions, and spatial resolutions. Using our dataset, we show that consumer LiDARs can enable accurate, data-driven NLOS perception. We further identify key scene and modeling factors that limit performance, as well as simulation-fidelity gaps that hinder current sim-to-real transfer, motivating future work toward scalable NLOS vision with consumer LiDARs.

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

DENALI releases a useful real-world dataset of low-cost LiDAR histograms for hidden objects, but the supporting experiments are too thin to back the accuracy claims.

read the letter

The main thing to know is that this paper introduces DENALI, a dataset of 72,000 real space-time histograms captured with consumer LiDARs across hidden-object scenes. That scale and the focus on actual hardware rather than simulation is the concrete advance here. Prior NLOS work has relied heavily on synthetic data or specialized sensors, so having diverse real captures with varied shapes, positions, lighting, and resolutions fills a gap that people working on practical robotics or mobile perception can use right away. The authors also call out limiting factors and sim-to-real mismatches, which keeps the contribution grounded instead of oversold. The dataset itself looks like something worth downloading and testing on. The soft spot is the missing evidence. The abstract states that the data enables accurate data-driven NLOS perception, yet it supplies no error metrics, no baselines, no ablations, and no indication of how well models extract the weak multi-bounce signals versus latching onto easier scene correlations. That leaves the central claim untested in the summary we have. The generalization concern is also reasonable: low-cost histograms are noisy and low-resolution, so models trained on this collection could easily fail outside the captured distribution even if in-distribution numbers look decent. This paper is for groups already working on computational imaging or low-cost robot sensing who need real training data or a benchmark. A reader focused on dataset-driven methods would get immediate value from the release, while someone expecting a full perception pipeline would find it preliminary. I would send it to peer review. The data collection is the load-bearing part and deserves referee scrutiny on capture protocol and suggested follow-up experiments, even if the modeling results need strengthening.

Referee Report

3 major / 2 minor

Summary. The paper introduces DENALI, the first large-scale real-world dataset of 72,000 space-time histograms captured from low-cost consumer LiDARs across diverse hidden-object scenes varying in shape, position, lighting, and resolution. It argues that the multi-bounce returns in these histograms enable data-driven NLOS perception of hidden objects, demonstrates this feasibility, and analyzes key limiting factors along with simulation-to-real gaps that hinder transfer.

Significance. If the empirical results hold, the dataset would be a valuable resource for NLOS research by shifting from simulation-only training to real captured histograms, potentially enabling practical hidden-object perception on commodity hardware in robotics and mobile devices. Explicitly identifying hardware constraints and sim-to-real discrepancies provides concrete directions for future model and sensor improvements.

major comments (3)

[Abstract] Abstract: The central claim that 'consumer LiDARs can enable accurate, data-driven NLOS perception' is asserted without any quantitative support such as accuracy, IoU, or error metrics, baselines (e.g., direct-return only or simulation-trained models), or test-set details. This absence makes it impossible to assess whether the learned mapping extracts usable multi-bounce signals or merely exploits dataset biases.
[Experiments] Experimental evaluation (assumed §4 or equivalent): No description of train/test splits, cross-scene generalization tests, or ablations isolating NLOS histogram components versus direct returns or scene correlations is provided. Without these, the skeptic concern that performance may collapse on unseen object placements or lighting cannot be evaluated.
[Dataset] Dataset capture description (assumed §3): The paper notes severe hardware limitations (low temporal resolution, single-photon noise) yet claims the histograms encode rich NLOS cues; however, no quantitative characterization of signal-to-noise ratios or multi-bounce visibility across the 72k scenes is given to ground this.

minor comments (2)

Clarify the exact LiDAR model, histogram bin count, and capture protocol (e.g., integration time, laser power) so that the dataset can be reproduced or extended by others.
The abstract mentions 'identifying key scene and modeling factors that limit performance' but does not list them explicitly; a table or enumerated list in the main text would improve readability.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their thoughtful and constructive review. We address each major comment below and agree that the manuscript would benefit from additional quantitative details and clarifications. We have revised the manuscript accordingly.

read point-by-point responses

Referee: [Abstract] Abstract: The central claim that 'consumer LiDARs can enable accurate, data-driven NLOS perception' is asserted without any quantitative support such as accuracy, IoU, or error metrics, baselines (e.g., direct-return only or simulation-trained models), or test-set details. This absence makes it impossible to assess whether the learned mapping extracts usable multi-bounce signals or merely exploits dataset biases.

Authors: We agree that the abstract would be strengthened by including quantitative support. The manuscript body demonstrates feasibility through experimental results, but we will revise the abstract to incorporate key metrics (accuracy, IoU, error rates), baseline comparisons, and test-set details to better substantiate the claim and allow immediate assessment of multi-bounce signal utility versus biases. revision: yes
Referee: [Experiments] Experimental evaluation (assumed §4 or equivalent): No description of train/test splits, cross-scene generalization tests, or ablations isolating NLOS histogram components versus direct returns or scene correlations is provided. Without these, the skeptic concern that performance may collapse on unseen object placements or lighting cannot be evaluated.

Authors: We acknowledge that more explicit experimental protocols are needed to evaluate generalization and rule out biases. While the manuscript includes results showing dataset utility for NLOS perception, we will expand the experimental section to detail train/test splits, add cross-scene generalization tests, and include ablations isolating NLOS multi-bounce components from direct returns and scene correlations. revision: yes
Referee: [Dataset] Dataset capture description (assumed §3): The paper notes severe hardware limitations (low temporal resolution, single-photon noise) yet claims the histograms encode rich NLOS cues; however, no quantitative characterization of signal-to-noise ratios or multi-bounce visibility across the 72k scenes is given to ground this.

Authors: We agree that quantitative characterization of SNR and multi-bounce visibility would better support the claims regarding NLOS cues despite the noted hardware constraints. We will add this analysis to the dataset capture section, including SNR measurements and visibility statistics across the 72,000 scenes. revision: yes

Circularity Check

0 steps flagged

Empirical dataset release with no derivation chain

full rationale

The paper's core contribution is the collection and release of 72,000 real-world space-time histogram scenes from low-cost LiDARs, followed by a feasibility demonstration that data-driven models can perform NLOS perception on this data. No equations, fitted parameters, uniqueness theorems, or predictive claims are advanced that could reduce to the inputs by construction. The abstract and described content contain no self-citation load-bearing steps, no ansatz smuggling, and no renaming of known results as novel derivations. The work is self-contained as an empirical benchmark and does not invoke any internal mathematical reduction that would trigger circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities are stated or required in the abstract; the work rests on the standard domain assumption that multi-bounce returns in LiDAR histograms encode usable NLOS information.

pith-pipeline@v0.9.0 · 5520 in / 1063 out tokens · 55970 ms · 2026-05-10T08:17:43.773666+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

41 extracted references · 5 canonical work pages · 1 internal anchor

[1]

TMF882X Datasheet.https:// ams-osram.com/, n.d

AMS OSRAM AG. TMF882X Datasheet.https:// ams-osram.com/, n.d. 2, 3, 4
[2]

ARKitScenes: A Diverse Real-World Dataset For 3D Indoor Scene Understanding Using Mobile RGB-D Data

Gilad Baruch, Zhuoyuan Chen, Afshin Dehghan, Tal Dimry, Yuri Feigin, Peter Fu, Thomas Gebauer, Brandon Joffe, Daniel Kurz, Arik Schwartz, et al. Arkitscenes: A diverse real-world dataset for 3d indoor scene understanding using mobile rgb-d data.arXiv preprint arXiv:2111.08897, 2021. 2

work page internal anchor Pith review arXiv 2021
[3]

Blurred lidar for sharper 3d: Robust handheld 3d scanning with diffuse lidar and rgb

Nikhil Behari, Aaron Young, Siddharth Somasundaram, Tzofi Klinghoffer, Akshat Dave, and Ramesh Raskar. Blurred lidar for sharper 3d: Robust handheld 3d scanning with diffuse lidar and rgb. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 26954– 26964, 2025. 2, 3

2025
[4]

Se- mantickitti: A dataset for semantic scene understanding of lidar sequences

Jens Behley, Martin Garbade, Andres Milioto, Jan Quen- zel, Sven Behnke, Cyrill Stachniss, and Jurgen Gall. Se- mantickitti: A dataset for semantic scene understanding of lidar sequences. InProceedings of the IEEE/CVF inter- national conference on computer vision, pages 9297–9307,
[5]

Lidar system architectures and cir- cuits.IEEE Communications Magazine, 55(10):135–142,

Behnam Behroozpour, Phillip AM Sandborn, Ming C Wu, and Bernhard E Boser. Lidar system architectures and cir- cuits.IEEE Communications Magazine, 55(10):135–142,
[6]

nuscenes: A multi- modal dataset for autonomous driving

Holger Caesar, Varun Bankiti, Alex H Lang, Sourabh V ora, Venice Erin Liong, Qiang Xu, Anush Krishnan, Yu Pan, Gi- ancarlo Baldan, and Oscar Beijbom. nuscenes: A multi- modal dataset for autonomous driving. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11621–11631, 2020. 2

2020
[7]

Low-cost spad sensing for non-line-of-sight tracking, material classification and depth imaging.ACM Transactions on Graphics (TOG), 40(4):1–12, 2021

Clara Callenberg, Zheng Shi, Felix Heide, and Matthias B Hullin. Low-cost spad sensing for non-line-of-sight tracking, material classification and depth imaging.ACM Transactions on Graphics (TOG), 40(4):1–12, 2021. 2, 3

2021
[8]

Learned feature embed- dings for non-line-of-sight imaging and recognition.ACM Transactions on Graphics (ToG), 39(6):1–18, 2020

Wenzheng Chen, Fangyin Wei, Kiriakos N Kutulakos, Szy- mon Rusinkiewicz, and Felix Heide. Learned feature embed- dings for non-line-of-sight imaging and recognition.ACM Transactions on Graphics (ToG), 39(6):1–18, 2020. 3

2020
[9]

Non-line-of-sight imaging.Nature Reviews Physics, 2(6): 318–327, 2020

Daniele Faccio, Andreas Velten, and Gordon Wetzstein. Non-line-of-sight imaging.Nature Reviews Physics, 2(6): 318–327, 2020. 3

2020
[10]

Nlos-neus: Non-line-of-sight neural im- plicit surface

Yuki Fujimura, Takahiro Kushida, Takuya Funatomi, and Ya- suhiro Mukaigawa. Nlos-neus: Non-line-of-sight neural im- plicit surface. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 10532–10541, 2023. 3

2023
[11]

Are we ready for autonomous driving? the kitti vision benchmark suite

Andreas Geiger, Philip Lenz, and Raquel Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In2012 IEEE conference on computer vision and pat- tern recognition, pages 3354–3361. IEEE, 2012. 2

2012
[12]

Multi-modal sensor fusion for auto driving perception: A survey,

Keli Huang, Botian Shi, Xiang Li, Xin Li, Siyuan Huang, and Yikang Li. Multi-modal sensor fusion for auto driv- ing perception: A survey.arXiv preprint arXiv:2202.02703,

work page arXiv
[13]

Depth Camera D435i Specifications

Intel RealSense. Depth Camera D435i Specifications. https : / / www . intel . com / content / www / us / en/products/sku/190004/intel- realsense- depth - camera - d435i / specifications . html, n.d. 4
[14]

Optical non-line-of-sight physics-based 3d human pose estimation

Mariko Isogawa, Ye Yuan, Matthew O’Toole, and Kris M Ki- tani. Optical non-line-of-sight physics-based 3d human pose estimation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 7013–7022,
[15]

Mitsuba 3 renderer, 2022.https://mitsuba-renderer.org

Wenzel Jakob, S ´ebastien Speierer, Nicolas Roussel, Merlin Nimier-David, Delio Vicini, Tizian Zeltner, Baptiste Nicolet, Miguel Crespo, Vincent Leroy, and Ziyi Zhang. Mitsuba 3 renderer, 2022.https://mitsuba-renderer.org. 5, 8

2022
[16]

Looking around the corner using transient imaging

Ahmed Kirmani, Tyler Hutchison, James Davis, and Ramesh Raskar. Looking around the corner using transient imaging. In2009 IEEE 12th International Conference on Computer Vision, pages 159–166. IEEE, 2009. 3

2009
[17]

Pointpillars: Fast encoders for object detection from point clouds

Alex H Lang, Sourabh V ora, Holger Caesar, Lubing Zhou, Jiong Yang, and Oscar Beijbom. Pointpillars: Fast encoders for object detection from point clouds. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12697–12705, 2019. 2

2019
[18]

Nlost: Non-line-of-sight imaging with transformer

Yue Li, Jiayong Peng, Juntian Ye, Yueyi Zhang, Feihu Xu, and Zhiwei Xiong. Nlost: Non-line-of-sight imaging with transformer. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13313– 13322, 2023. 3

2023
[19]

Toward dynamic non-line-of- sight imaging with mamba enforced temporal consistency

Yue Li, Yi Sun, Shida Sun, Juntian Ye, Yueyi Zhang, Feihu Xu, and Zhiwei Xiong. Toward dynamic non-line-of- sight imaging with mamba enforced temporal consistency. Advances in Neural Information Processing Systems, 37: 126452–126473, 2024. 3

2024
[20]

Hand- held mapping of specular surfaces using consumer-grade flash lidar

Tsung-Han Lin, Connor Henley, Siddharth Somasundaram, Akshat Dave, Moshe Laifenfeld, and Ramesh Raskar. Hand- held mapping of specular surfaces using consumer-grade flash lidar. In2024 IEEE International Conference on Com- putational Photography (ICCP), pages 1–12. IEEE, 2024. 2

2024
[21]

Towards 3d vision with low-cost single-photon cameras

Fangzhou Mu, Carter Sifferman, Sacha Jungerman, Yiquan Li, Mark Han, Michael Gleicher, Mohit Gupta, and Yin Li. Towards 3d vision with low-cost single-photon cameras. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 5302–5311, 2024. 2, 3

2024
[22]

A light transport model for mitigating multipath interference in time-of-flight sensors

Nikhil Naik, Achuta Kadambi, Christoph Rhemann, Shahram Izadi, Ramesh Raskar, and Sing Bing Kang. A light transport model for mitigating multipath interference in time-of-flight sensors. InProceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition, pages 73–81, 2015. 2

2015
[23]

A review of single-photon avalanche diode time-of-flight imaging sensor arrays.IEEE Sensors Journal, 21(11):12654–12666, 2020

Franc ¸ois Piron, Daniel Morrison, Mehmet Rasit Yuce, and Jean-Michel Redout´e. A review of single-photon avalanche diode time-of-flight imaging sensor arrays.IEEE Sensors Journal, 21(11):12654–12666, 2020. 2

2020
[24]

Computational imaging based on single-photon detection: a survey.Artificial Intelli- gence Review, 58(8):251, 2025

Yanyun Pu, Chengyuan Zhu, Gongxin Yao, Chao Li, Yu Pan, Kaixiang Yang, and Qinmin Yang. Computational imaging based on single-photon detection: a survey.Artificial Intelli- gence Review, 58(8):251, 2025. 2

2025
[25]

3d object detection for autonomous driving: A survey.Pattern Recognition, 130: 108796, 2022

Rui Qian, Xin Lai, and Xirong Li. 3d object detection for autonomous driving: A survey.Pattern Recognition, 130: 108796, 2022. 2

2022
[26]

Automotive lidar technology: A survey.IEEE Transactions on Intelligent Transportation Systems, 23(7):6282–6297, 2021

Ricardo Roriz, Jorge Cabral, and Tiago Gomes. Automotive lidar technology: A survey.IEEE Transactions on Intelligent Transportation Systems, 23(7):6282–6297, 2021. 2

2021
[27]

mitran- sient: Transient light transport in mitsuba 3.arXiv preprint arXiv:2510.25660, 2025

Diego Royo, Jorge Garcia-Pueyo, Miguel Crespo, ´Oscar Pueyo-Ciutad, Guillermo Enguita, and Diego Bielsa. mitran- sient: Transient light transport in mitsuba 3.arXiv preprint arXiv:2510.25660, 2025. 8

work page arXiv 2025
[28]

All photons imaging through volumetric scattering

Guy Satat, Barmak Heshmat, Dan Raviv, and Ramesh Raskar. All photons imaging through volumetric scattering. Scientific reports, 6(1):33946, 2016. 2

2016
[29]

Towards photography through realistic fog

Guy Satat, Matthew Tancik, and Ramesh Raskar. Towards photography through realistic fog. In2018 IEEE Interna- tional Conference on Computational Photography (ICCP), pages 1–10. IEEE, 2018. 2

2018
[30]

Non-line-of- sight imaging via neural transient fields.IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(7):2257– 2268, 2021

Siyuan Shen, Zi Wang, Ping Liu, Zhengqing Pan, Ruiqian Li, Tian Gao, Shiying Li, and Jingyi Yu. Non-line-of- sight imaging via neural transient fields.IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(7):2257– 2268, 2021. 3

2021
[31]

Effi- cient detection of objects near a robot manipulator via minia- ture time-of-flight sensors.IEEE Robotics and Automation Letters, 2025

Carter Sifferman, Mohit Gupta, and Michael Gleicher. Effi- cient detection of objects near a robot manipulator via minia- ture time-of-flight sensors.IEEE Robotics and Automation Letters, 2025. 3

2025
[32]

Recovering parametric scenes from very few time-of-flight pixels

Carter Sifferman, Yiquan Li, Yiming Li, Fangzhou Mu, Michael Gleicher, Mohit Gupta, and Yin Li. Recovering parametric scenes from very few time-of-flight pixels. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 27989–27999, 2025. 3

2025
[33]

VL53L8CX Product Overview

STMicroelectronics. VL53L8CX Product Overview. https : / / www . st . com / en / imaging - and - photonics-solutions/vl53l8cx.html, n.d. 2, 3
[34]

Material classification using raw time-of-flight measure- ments

Shuochen Su, Felix Heide, Robin Swanson, Jonathan Klein, Clara Callenberg, Matthias Hullin, and Wolfgang Heidrich. Material classification using raw time-of-flight measure- ments. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 3503–3511, 2016. 2

2016
[35]

Scalability in perception for autonomous driving: Waymo open dataset

Pei Sun, Henrik Kretzschmar, Xerxes Dotiwalla, Aurelien Chouard, Vijaysai Patnaik, Paul Tsui, James Guo, Yin Zhou, Yuning Chai, Benjamin Caine, et al. Scalability in perception for autonomous driving: Waymo open dataset. InProceed- ings of the IEEE/CVF conference on computer vision and pattern recognition, pages 2446–2454, 2020. 2

2020
[36]

Data-driven non-line-of-sight imaging with a tradi- tional camera

Matthew Tancik, Tristan Swedish, Guy Satat, and Ramesh Raskar. Data-driven non-line-of-sight imaging with a tradi- tional camera. InImaging Systems and Applications, pages IW2B–6. Optica Publishing Group, 2018. 3

2018
[37]

Apriltag: Detection and pose estimation li- brary for c, c++ and python.https://github.com/ Tinker-Twins/AprilTag, 2023

Tinker-Twins. Apriltag: Detection and pose estimation li- brary for c, c++ and python.https://github.com/ Tinker-Twins/AprilTag, 2023. Accessed: 2025-11-

2023
[38]

Recovering three-dimensional shape around a corner using ultrafast time-of-flight imaging.Nature communications, 3 (1):745, 2012

Andreas Velten, Thomas Willwacher, Otkrist Gupta, Ashok Veeraraghavan, Moungi G Bawendi, and Ramesh Raskar. Recovering three-dimensional shape around a corner using ultrafast time-of-flight imaging.Nature communications, 3 (1):745, 2012. 3

2012
[39]

arXiv preprint arXiv:2008.08063 (2020)

Xinshuo Weng, Jianren Wang, David Held, and Kris Kitani. Ab3dmot: A baseline for 3d multi-object tracking and new evaluation metrics.arXiv preprint arXiv:2008.08063, 2020. 2

work page arXiv 2008
[40]

Batagoda, Harry Zhang, Akshat Dave, Adithya Pediredla, Dan Negrut, and Ramesh Raskar

Aaron Young, Nevindu M. Batagoda, Harry Zhang, Akshat Dave, Adithya Pediredla, Dan Negrut, and Ramesh Raskar. Enhancing Autonomous Navigation by Imaging Hidden Ob- jects using Single-Photon LiDAR, 2025. arXiv:2410.03555 [cs]. 3

work page arXiv 2025
[41]

Lidar-based slam for robotic mapping: state of the art and new frontiers.Industrial Robot: the international journal of robotics research and applica- tion, 51(2):196–205, 2024

Xiangdi Yue, Yihuan Zhang, Jiawei Chen, Junxin Chen, Xu- anyi Zhou, and Miaolei He. Lidar-based slam for robotic mapping: state of the art and new frontiers.Industrial Robot: the international journal of robotics research and applica- tion, 51(2):196–205, 2024. 2

2024