pith. sign in

arxiv: 2604.11406 · v2 · submitted 2026-04-13 · 💻 cs.RO

Using Unwrapped Full Color Space Recording to Measure the Exposedness of Vehicle Exterior Parts for External Human Machine Interfaces

Jose Gonzalez-Belmonte , Jaerock Kwon This is my paper

Pith reviewed 2026-05-10 15:56 UTC · model grok-4.3

classification 💻 cs.RO
keywords autonomous vehiclesexternal human-machine interfacespedestrian visibilityvehicle exterior placementsimulation measurementintent communicationintersection crossing
0
0 comments X

The pith

A simulation of pedestrian crossings finds that eHMIs on bumpers and grills can be blocked by other vehicles, so visibility improves with simultaneous placement on the windshield and frontal fenders.

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

The paper introduces unwrapped full color space recording on a vehicle mesh inside Unity to count how often each exterior point remains visible to a crossing pedestrian at a four-way intersection. It reports that the bumper, grill, and hood appear most often in the pedestrian's view, yet nearby cars on the same lane frequently obstruct those surfaces. The work therefore concludes that a single-location design is unreliable and instead recommends placing the interfaces on both the windshield and frontal fenders at the same time. A sympathetic reader would care because autonomous vehicles need dependable ways to signal intent to pedestrians if collisions are to be avoided. The color-mapping technique turns every mesh face into a unique identifier so that visibility can be tallied frame by frame without manual inspection.

Core claim

By unwrapping the exterior mesh of a 2015 Ford F-150 into a full-color texture and running animated pedestrian-crossing scenes in Unity, the study records which colored regions are unobstructed from the pedestrian's viewpoint across many frames. Cross-referencing these counts with a part-labeled map shows the bumper, grill, and hood as the most exposed surfaces, but the presence of other vehicles in the same lane can hide them. The authors therefore conclude that eHMI placement must be distributed, specifically combining the windshield and frontal fenders, to maintain reliable visibility regardless of surrounding traffic.

What carries the argument

Unwrapped full color space recording, which assigns each face of the vehicle mesh a unique color on a texture map so that simulation frames can tally visible parts by color detection.

If this is right

  • Bumper, grill, and hood placements alone are insufficient when other cars occupy the same lane.
  • Simultaneous use of windshield and frontal fenders supplies more consistent visibility to a crossing pedestrian.
  • The color-unwrapping method can be reapplied to test visibility on different vehicle shapes or intersection geometries.
  • Reliable eHMI placement supports clearer intent signaling and may reduce pedestrian hesitation around autonomous vehicles.

Where Pith is reading between the lines

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

  • The same recording technique could be extended to measure visibility for cyclists or drivers in adjacent lanes.
  • Design guidelines for autonomous vehicles might eventually specify minimum exposed area requirements based on this kind of simulation data.
  • Physical experiments that vary lighting, weather, and pedestrian height would test whether the simulation rankings remain stable.

Load-bearing premise

The chosen Unity simulation of one pedestrian crossing a four-way intersection captures the same visibility conditions that would occur with real vehicles, real pedestrians, and varied traffic layouts.

What would settle it

A side-by-side comparison of the simulation counts against footage from cameras mounted on an actual Ford F-150 at a real intersection with multiple vehicles present.

Figures

Figures reproduced from arXiv: 2604.11406 by Jaerock Kwon, Jose Gonzalez-Belmonte.

Figure 2
Figure 2. Figure 2: A rendition of Full Color Space Palette (FCSP) generated for this [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Faces of the F-150 King Ranch mesh unwrapped onto the FCSP [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The vehicle model in the 3D engine with the FCSP assigned as its [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: Compressed visualization of a 6, 420 × 8, 100 image being divided into twenty-five 1, 284 × 1, 620 files before capturing and saving. During the recording of an animation, a script advances the animation for the selected scenario a specified amount of time, and has both binocular cameras alternate taking a capture of the frame at that time. F. Recording Data by Vehicle Exterior Parts One of the limitations… view at source ↗
Figure 7
Figure 7. Figure 7: Exterior elements of F150 King Rach uniquely identified by color. [PITH_FULL_IMAGE:figures/full_fig_p005_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Unwrapped texture of the Ford F-150 King Ranch with each of its [PITH_FULL_IMAGE:figures/full_fig_p006_8.png] view at source ↗
Figure 10
Figure 10. Figure 10: Pictures showing the exposedness of the exterior of the vehicle [PITH_FULL_IMAGE:figures/full_fig_p007_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Pictures showing the exposedness of the exterior of the vehicle [PITH_FULL_IMAGE:figures/full_fig_p007_11.png] view at source ↗
read the original abstract

One of the concerns with autonomous vehicles is their ability to communicate their intent to other road users, specially pedestrians, in order to prevent accidents. External Human-Machine Interfaces (eHMIs) are the proposed solution to this issue, through the introduction of electronic devices on the exterior of a vehicle that communicate when the vehicle is planning on slowing down or yielding. This paper uses the technique of unwrapping the faces of a mesh onto a texture where every pixel is a unique color, as well as a series of animated simulations made and ran in the Unity game engine, to measure how many times is each point on a 2015 Ford F-150 King Ranch is unobstructed to a pedestrian attempting to cross the road at a four-way intersection. By cross-referencing the results with a color-coded map of the labeled parts on the exterior of the vehicle, it was concluded that while the bumper, grill, and hood were the parts of the vehicle visible to the crossing pedestrian most often, the existence of other vehicles on the same lane that might obstruct the view of these makes them insufficient. The study recommends instead a distributive approach to eHMIs by using both the windshield and frontal fenders as simultaneous placements for these devices.

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.

Referee Report

2 major / 2 minor

Summary. The paper describes a Unity-based simulation method that unwraps a 2015 Ford F-150 mesh onto a unique-color texture to count visible pixels of each exterior part from a single pedestrian viewpoint at a four-way intersection. It reports that the bumper, grill, and hood are the most frequently visible parts but concludes these are insufficient because other vehicles in the same lane may occlude them, and therefore recommends a distributive eHMI placement using both the windshield and frontal fenders simultaneously.

Significance. If the visibility rankings prove robust, the work supplies a quantitative, mesh-unwrapping technique for measuring exposedness that could guide eHMI placement decisions on autonomous vehicles. The approach of assigning unique colors to mesh faces and tallying rendered pixels offers a precise, repeatable way to compare part visibility without manual annotation, which is a methodological strength.

major comments (2)
  1. [Results and Discussion] The central recommendation that bumper/grill/hood placements are insufficient (and that windshield + frontal fenders should be used instead) rests entirely on pixel-count rankings obtained from one static pedestrian position and one four-way intersection configuration that includes other vehicles. No sensitivity analysis to pedestrian height, lateral position, or traffic density is reported, so the occlusion argument cannot be assessed for robustness.
  2. [Methodology] The simulation logic is described as straightforward, yet the manuscript supplies no validation of the Unity rendering against real-world camera data, no error bounds on the color-unwrapping counts, and no comparison to physical line-of-sight measurements. Because the part-ranking and placement advice depend directly on these counts being representative, the absence of such checks is load-bearing.
minor comments (2)
  1. [Abstract] The abstract contains minor grammatical issues (e.g., 'specially' should be 'especially'; 'how many times is each point ... unobstructed' is awkward) that should be corrected for clarity.
  2. [Introduction] The paper would benefit from explicit citation of prior eHMI visibility studies to situate the quantitative method against existing qualitative or camera-based approaches.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for the constructive comments, which help clarify the scope of our simulation study on measuring vehicle part visibility for eHMI design. We respond to each major comment below.

read point-by-point responses

  1. Referee: [Results and Discussion] The central recommendation that bumper/grill/hood placements are insufficient (and that windshield + frontal fenders should be used instead) rests entirely on pixel-count rankings obtained from one static pedestrian position and one four-way intersection configuration that includes other vehicles. No sensitivity analysis to pedestrian height, lateral position, or traffic density is reported, so the occlusion argument cannot be assessed for robustness.

    Authors: We agree that the analysis uses a single pedestrian viewpoint and intersection layout. This specific setup was selected to illustrate a common crossing scenario in which other vehicles in the same lane can occlude the front bumper, grill, and hood. The distributive placement recommendation follows directly from that observation in the simulated case. In the revised manuscript we will add explicit discussion of this limitation, stating that the findings are illustrative rather than universal and that sensitivity analyses across pedestrian heights, positions, and traffic densities would be valuable extensions. revision: partial

  2. Referee: [Methodology] The simulation logic is described as straightforward, yet the manuscript supplies no validation of the Unity rendering against real-world camera data, no error bounds on the color-unwrapping counts, and no comparison to physical line-of-sight measurements. Because the part-ranking and placement advice depend directly on these counts being representative, the absence of such checks is load-bearing.

    Authors: The manuscript introduces the color-unwrapping technique as a deterministic, annotation-free method for counting visible pixels inside a Unity simulation. Because the work is simulation-only, real-world camera validation and physical line-of-sight comparisons were outside its scope. We will revise the methodology and limitations sections to state the rendering assumptions, note that pixel counts are exact within the simulated environment, and acknowledge that discrepancies with real cameras (lighting, lens effects) remain unquantified. This will make the intended scope of the rankings clearer. revision: yes

standing simulated objections not resolved
  • Real-world validation of the Unity rendering against camera data and physical line-of-sight measurements

Circularity Check

0 steps flagged

No circularity: results are direct simulation pixel counts

full rationale

The paper derives part-visibility rankings solely from Unity rendering counts after assigning unique colors to unwrapped mesh faces and tallying unobstructed pixels from a fixed pedestrian viewpoint. No equations, fitted parameters, or predictions are present that reduce to their own inputs by construction. No self-citations, uniqueness theorems, or ansatzes are invoked as load-bearing steps. The method is a straightforward measurement pipeline whose outputs (visibility frequencies per part) follow directly from the simulation geometry and color-mapping procedure without circular reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no information on free parameters, axioms, or invented entities.

pith-pipeline@v0.9.0 · 5524 in / 1048 out tokens · 51422 ms · 2026-05-10T15:56:00.499302+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

31 extracted references · 31 canonical work pages · 1 internal anchor

  1. [1]

    Holiday, 2015 Ford F150 King Ranch Edition

    D. Holiday, 2015 Ford F150 King Ranch Edition

    work page 2015

  2. [2]

    26, 2026

    Accessed: Mar. 26, 2026. [3D Model]. Available: https://sketchfab.com/models/79ebd5caaf414ab08debc36157dbc878/

    work page 2026

  3. [3]

    A Novel Way of Optimizing Headlight Distributions Based on Real Life Traffic and Eye Tracking Data Part 1: Idealized Baseline Distribution,

    J. Kobbert, A. Erkan, J. D. Bullough, and T. Q. Khanh, “A Novel Way of Optimizing Headlight Distributions Based on Real Life Traffic and Eye Tracking Data Part 1: Idealized Baseline Distribution,” Applied Sciences, vol. 13, no. 17, p. 9908, Jan. 2023, doi: 10.3390/app13179908

    work page doi:10.3390/app13179908 2023

  4. [4]

    Accessed: Mar

    Alexander Christiaan Jacob, “allRGB,” allRGB. Accessed: Mar. 26,

    work page

  5. [5]

    Available: https://allrgb.com/

    [Online]. Available: https://allrgb.com/

    work page

  6. [6]

    London: Henry Kimpton,

    Harry Moss Traquair, An Introduction to Clinical Perimetry, Sixth Edition. London: Henry Kimpton,

    work page

  7. [7]

    26, 2026

    Accessed: Mar. 26, 2026. [Online]. Available: https://ia802902.us.archive.org/1/items/in.ernet.dli.2015.146333/2015.146333.An- Introduction-To-Clinical-Perimetry-Sixth-Edition.pdf

    work page arXiv 2026

  8. [8]

    Analytical Study on the Exposedness of Potential Positions for External Human-Machine Interfaces

    J. Gonzalez-Belmonte and J. Kwon, “Analytical Study on the Exposedness of Potential Positions for External Human- Machine Interfaces,” Feb. 13, 2026, arXiv: arXiv:2507.08973. doi: 10.48550/arXiv.2507.08973.. Unpublished

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2507.08973 2026

  9. [9]

    Blender,

    Blender Foundation, “Blender,” Blender. Accessed: Mar. 26, 2026. [Online]. Available: https://www.blender.org/

    work page 2026

  10. [10]

    Corrections to: Peripheral vision and pattern recognition: A review,

    H. Strasburger, I. Rentschler, and M. J ¨uttner, “Corrections to: Peripheral vision and pattern recognition: A review,” Journal of Vision, vol. 24, no. 4, p. 15, Apr. 2024, doi: 10.1167/jov.24.4.15

    work page doi:10.1167/jov.24.4.15 2024

  11. [11]

    Effect of eHMI-equipped automated vehicles on pedestrian crossing behavior and safety: A focus on blind spot scenarios,

    X. Chen, X. Li, Y . Hou, W. Yang, C. Dong, and H. Wang, “Effect of eHMI-equipped automated vehicles on pedestrian crossing behavior and safety: A focus on blind spot scenarios,” Accident Analysis & Preven- tion, vol. 212, p. 107915, Mar. 2025, doi: 10.1016/j.aap.2025.107915

    work page doi:10.1016/j.aap.2025.107915 2025

  12. [12]

    , author Boissieux, L

    M. Troel–madec et al., “eHMI positioning for autonomous vehi- cle/pedestrians interaction,” in IHM 2019 - 31e conf ´erence francophone sur l’Interaction Homme-Machine, Grenoble, France: ACM, Dec. 2019, pp. 1–8. doi: 10.1145/3366551.3370340

    work page doi:10.1145/3366551.3370340 2019

  13. [13]

    , author Dietrich, A

    K. de Clercq, A. Dietrich, J. P. N ´u˜nez Velasco, J. de Winter, and R. Happee, “External Human-Machine Interfaces on Automated Vehicles: Effects on Pedestrian Crossing Decisions,” Hum Factors, vol. 61, no. 8, pp. 1353–1370, Dec. 2019, doi: 10.1177/0018720819836343

    work page doi:10.1177/0018720819836343 2019

  14. [14]

    , author Dodou, D

    J. de Winter and D. Dodou, “External human–machine interfaces: Gim- mick or necessity?,” Transportation Research Interdisciplinary Perspec- tives, vol. 15, p. 100643, Sep. 2022, doi: 10.1016/j.trip.2022.100643

    work page doi:10.1016/j.trip.2022.100643 2022

  15. [15]

    From Visions to Reality: Investigating the Interplay of Vehicle Kinematics and Light- band eHMI in a Real Vehicle Study,

    M. Lau, H. P. Nguyen, M. Jipp, and M. Oehl, “From Visions to Reality: Investigating the Interplay of Vehicle Kinematics and Light- band eHMI in a Real Vehicle Study,” Transportation Research Part F: Traffic Psychology and Behaviour, vol. 103, pp. 79–95, May 2024, doi: 10.1016/j.trf.2024.03.004

    work page doi:10.1016/j.trf.2024.03.004 2024

  16. [16]

    , author van der Kint, S

    A. Rodr ´ıguez Palmeiro, S. van der Kint, L. Vissers, H. Farah, J. C. F. de Winter, and M. Hagenzieker, “Interaction between pedestrians and auto- mated vehicles: A Wizard of Oz experiment,” Transportation Research Part F: Traffic Psychology and Behaviour, vol. 58, pp. 1005–1020, Oct. 2018, doi: 10.1016/j.trf.2018.07.020

    work page doi:10.1016/j.trf.2018.07.020 2018

  17. [17]

    Normal values and standard deviations for pupil diameter and interpupillary distance in subjects aged 1 month to 19 years,

    C. MacLachlan and H. C. Howland, “Normal values and standard deviations for pupil diameter and interpupillary distance in subjects aged 1 month to 19 years,” Ophthalmic Physiol Opt, vol. 22, no. 3, pp. 175–182, May 2002, doi: 10.1046/j.1475-1313.2002.00023.x

    work page doi:10.1046/j.1475-1313.2002.00023.x 2002

  18. [18]

    Most Popular Cars in America,

    “Most Popular Cars in America,” Edmunds. Accessed: Mar. 26, 2026. [Online]. Available: https://www.edmunds.com/most-popular-cars/

    work page 2026

  19. [19]

    Measured average height, weight, and waist circumference for adults ages 20 and older

    National Center for Health Statistics, “Measured average height, weight, and waist circumference for adults ages 20 and older.” Accessed: Apr. 06, 2026. [Online]. Available: https://www.cdc.gov/nchs/fastats/body- measurements.htm

    work page 2026

  20. [20]

    Median Age in 192 Metro Areas Higher Than National Median of 39.1,

    K. Wilder and P. Mackun, “Median Age in 192 Metro Areas Higher Than National Median of 39.1,” Census.gov. Accessed: Apr. 06, 2026. [On- line]. Available: https://www.census.gov/library/stories/2025/06/metro- areas-median-age.html

    work page 2026

  21. [21]

    Peripheral vision and pattern recognition: A review,

    H. Strasburger, I. Rentschler, and M. J ¨uttner, “Peripheral vision and pattern recognition: A review,” Journal of Vision, vol. 11, no. 5, p. 13, Dec. 2011, doi: 10.1167/11.5.13

    work page doi:10.1167/11.5.13 2011

  22. [22]

    Resolution limit of the eye — how many pixels can we see?,

    M. Ashraf, A. Chapiro, and R. K. Mantiuk, “Resolution limit of the eye — how many pixels can we see?,” Nat Commun, vol. 16, no. 1, p. 9086, Oct. 2025, doi: 10.1038/s41467-025-64679-2

    work page doi:10.1038/s41467-025-64679-2 2025

  23. [23]

    Auberbach and D

    T. Auberbach and D. E. Kelm, RGB Colorspace Atlas. Accessed: Mar. 26, 2026. [Online]. Available: https://taubaauerbach.com/view.php?id=286

    work page 2026

  24. [24]

    Seven Myths on Crowding and Peripheral Vision,

    H. Strasburger, “Seven Myths on Crowding and Peripheral Vision,” i-Perception, vol. 11, no. 3, p. 2041669520913052, May 2020, doi: 10.1177/2041669520913052

    work page doi:10.1177/2041669520913052 2020

  25. [25]

    Study the Effect of eHMI Projection Distance and Contrast on People Acceptance in Blind-Spot Detection Scenario,

    A. H. Shah, X. Sun, and Y . Lin, “Study the Effect of eHMI Projection Distance and Contrast on People Acceptance in Blind-Spot Detection Scenario,” Applied Sciences, vol. 12, no. 13, p. 6730, Jan. 2022, doi: 10.3390/app12136730

    work page doi:10.3390/app12136730 2022

  26. [26]

    , author Dodou, D

    P. Bazilinskyy, D. Dodou, and J. de Winter, “Survey on eHMI concepts: The effect of text, color, and perspective,” Transportation Research Part F: Traffic Psychology and Behaviour, vol. 67, pp. 175–194, Nov. 2019, doi: 10.1016/j.trf.2019.10.013

    work page doi:10.1016/j.trf.2019.10.013 2019

  27. [27]

    The effect of drivers’ eye contact on pedestrians’ perceived safety,

    V . Onkhar, P. Bazilinskyy, D. Dodou, and J. C. F. de Winter, “The effect of drivers’ eye contact on pedestrians’ perceived safety,” Trans- portation Research Part F: Traffic Psychology and Behaviour, vol. 84, pp. 194–210, Jan. 2022, doi: 10.1016/j.trf.2021.10.017

    work page doi:10.1016/j.trf.2021.10.017 2022

  28. [28]

    , author Kwigizile, V

    A. Alhawiti et al., “The Effectiveness of eHMI Displays on Pedestrian–Autonomous Vehicle Interaction in Mixed-Traffic Environ- ments,” Sensors (Basel), vol. 24, no. 15, p. 5018, Aug. 2024, doi: 10.3390/s24155018

    work page doi:10.3390/s24155018 2024

  29. [29]

    The Most Popular Cars in America,

    A. Vidgerrman, “The Most Popular Cars in America,” The Most Popular Cars in America. Accessed: Mar. 26, 2026. [Online]. Available: https://www.autoinsurance.com/articles/most-popular-cars-america/

    work page 2026

  30. [30]

    Accessed: Mar

    “Unity,” Unity. Accessed: Mar. 26, 2026. [Online]. Available: https://unity.com

    work page 2026

  31. [31]

    What are PNG files and how do you open them? — Adobe

    “What are PNG files and how do you open them? — Adobe.” Accessed: Apr. 06, 2026. [Online]. Available: https://www.adobe.com/creativecloud/file-types/image/raster/png- file.html

    work page 2026