pith. sign in

arxiv: 2404.17104 · v1 · submitted 2024-04-26 · 💻 cs.HC · cs.CV

Don't Look at the Camera: Achieving Perceived Eye Contact

Alice Gao , Samyukta Jayakumar , Marcello Maniglia , Brian Curless , Ira Kemelmacher-Shlizerman , Aaron R. Seitz , Steven M. Seitz This is my paper

Pith reviewed 2026-05-24 02:30 UTC · model grok-4.3

classification 💻 cs.HC cs.CV
keywords eye contact perceptiongaze directioncamera lens offsetuser studiesvideo communication2D display rendering
0
0 comments X

The pith

To create the perception of eye contact on a 2D screen, subjects should gaze just below the camera lens rather than directly into it.

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

Standard guidance holds that looking straight into the camera produces the best sense of eye contact when video or photos are shown on a flat display. Controlled user studies instead find that directing the gaze a small distance below the lens yields higher ratings for perceived eye contact. The work measures this effect quantitatively by varying the vertical offset of the subject's gaze and collecting viewer judgments. The result applies to single subjects under typical capture and viewing conditions. It supplies a practical adjustment for anyone appearing in video calls, recorded talks, or still images.

Core claim

Through empirical user studies, we show that it is instead preferable to look just below the camera lens to optimize the perception that they are making eye contact. We quantitatively assess where subjects should direct their gaze relative to a camera lens to optimize the perception that they are making eye contact.

What carries the argument

User studies that rate perceived eye contact while systematically varying the subject's vertical gaze offset relative to the camera lens.

If this is right

  • Participants in video calls can be directed to look slightly below the camera to increase the sense of eye contact.
  • Camera placement and subject instructions in photography and videography may shift away from direct lens gaze.
  • The optimal vertical offset can be measured and applied as a standard adjustment for single-subject 2D captures.
  • The finding is specific to flat displays and does not address multi-person or 3D rendering cases.

Where Pith is reading between the lines

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

  • Device makers could embed visual cues or software prompts that guide users to the below-lens position during recording.
  • The same offset principle might interact with head angle or facial expression in ways not tested here.
  • Remote education or therapy sessions could adopt the adjustment to strengthen perceived connection between speaker and viewer.

Load-bearing premise

The specific user studies and viewing conditions used accurately capture general human perception of eye contact and generalize beyond the tested setups and participant groups.

What would settle it

A replication with new participants, different screen sizes, or changed viewing distances that finds direct lens gaze receives higher eye-contact ratings than the below-lens position.

Figures

Figures reproduced from arXiv: 2404.17104 by Aaron R. Seitz, Alice Gao, Brian Curless, Ira Kemelmacher-Shlizerman, Marcello Maniglia, Samyukta Jayakumar, Steven M. Seitz.

Figure 1
Figure 1. Figure 1: (left) Directing our gaze at a camera lens is perceived as looking slightly upwards. (right) Looking two [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Equipment set up used to collect pictures from gazers. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Single trial during data collection for reporting subjective gaze perception. All trials were self-paced, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: (a) Average distribution of ‘yes’ response across gaze offset with 95% CI and Gaussian fit showing [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
read the original abstract

We consider the question of how to best achieve the perception of eye contact when a person is captured by camera and then rendered on a 2D display. For single subjects photographed by a camera, conventional wisdom tells us that looking directly into the camera achieves eye contact. Through empirical user studies, we show that it is instead preferable to {\em look just below the camera lens}. We quantitatively assess where subjects should direct their gaze relative to a camera lens to optimize the perception that they are making eye contact.

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

1 major / 0 minor

Summary. The paper claims that conventional advice to look directly into the camera for perceived eye contact on 2D displays is suboptimal. Through empirical user studies, it instead finds that directing gaze just below the camera lens optimizes the perception of eye contact, and it quantitatively assesses the preferred angular offset relative to the lens.

Significance. If the user studies are robust, this result could inform practical guidelines for video conferencing setups and camera placement in HCI systems. The work is purely empirical with no free parameters or derivations, so its value hinges on the quality and generalizability of the experiments.

major comments (1)
  1. [Abstract] Abstract and methods description: no information is supplied on participant count, demographics, statistical tests, exact angular offsets tested, display size, camera-to-face geometry, or controls for confounds such as lighting and vertical camera offset. This directly undermines assessment of whether the studies support the central claim or generalize beyond the tested setups, as highlighted by the stress-test concern.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their review and for highlighting the need for greater methodological transparency. We address the single major comment below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract and methods description: no information is supplied on participant count, demographics, statistical tests, exact angular offsets tested, display size, camera-to-face geometry, or controls for confounds such as lighting and vertical camera offset. This directly undermines assessment of whether the studies support the central claim or generalize beyond the tested setups, as highlighted by the stress-test concern.

    Authors: We agree that the abstract omits these details and that their absence limits evaluation of the work. In the revised manuscript we will expand the abstract to report participant count, demographics, statistical tests, exact angular offsets, display size, camera-to-face geometry, and controls for lighting and vertical offset. The methods section will be clarified and cross-referenced so that readers can fully assess robustness and generalizability. revision: yes

Circularity Check

0 steps flagged

No circularity: purely empirical user-study paper

full rationale

The paper's central claim is established through empirical user studies comparing gaze directions relative to a camera lens, with no equations, derivations, fitted parameters, or self-citation chains invoked to support the result. The abstract and described approach contain no load-bearing steps that reduce to self-definition, renaming, or imported uniqueness theorems. The derivation chain is therefore self-contained against external benchmarks (the studies themselves), warranting a score of 0.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No mathematical model or new entities; the work rests on standard assumptions of perceptual psychology and experimental design in HCI.

pith-pipeline@v0.9.0 · 5629 in / 906 out tokens · 17816 ms · 2026-05-24T02:30:11.358432+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

26 extracted references · 26 canonical work pages

  1. [1]

    Sean Andrist, Michael Gleicher, and Bilge Mutlu. 2017. Looking Coordinated: Bidirectional Gaze Mechanisms for Collaborative Interaction with Virtual Characters. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems (Denver, Colorado, USA) (CHI ’17). Association for Computing Machinery, New York, NY, USA, 2571–2582. https://doi.or...

    work page doi:10.1145/3025453.3026033 2017

  2. [2]

    Stuart M Anstis, John W Mayhew, and Tania Morley. 1969. The perception of where a face or television’portrait’is looking. The American journal of psychology 82, 4 (1969), 474–489

    work page 1969

  3. [3]

    Michael Argyle and Mark Cook. 1976. Gaze and mutual gaze. (1976)

    work page 1976

  4. [4]

    Simon Baron-Cohen, Ruth Campbell, Annette Karmiloff-Smith, Julia Grant, and Jane Walker. 1995. Are children with autism blind to the mentalistic significance of the eyes? British Journal of Developmental Psychology 13, 4 (1995), 379–398

    work page 1995

  5. [5]

    Andrew P Bayliss, Jessica Bartlett, Claire K Naughtin, and Ada Kritikos. 2011. A direct link between gaze perception and social attention. Journal of Experimental Psychology: Human Perception and Performance 37, 3 (2011), 634

    work page 2011

  6. [6]

    David H Brainard and Spatial Vision. 1997. The psychophysics toolbox. Spatial vision 10, 4 (1997), 433–436

    work page 1997

  7. [7]

    Andrew J Calder. 2003. Disgust discussed. Annals of Neurology (2003). 8 Gao, Jayakumar, Maniglia, et al

    work page 2003

  8. [8]

    Milton Chen. 2002. Leveraging the Asymmetric Sensitivity of Eye Contact for Videoconference. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (Minneapolis, Minnesota, USA) (CHI ’02). Association for Computing Machinery, New York, NY, USA, 49–56. https://doi.org/10.1145/503376.503386

    work page doi:10.1145/503376.503386 2002

  9. [9]

    Marvin G Cline. 1967. The perception of where a person is looking. The American journal of psychology 80, 1 (1967), 41–50

    work page 1967

  10. [10]

    Caroline Gale and Andrew Monk. 2012. Where am I looking? The accuracy of video-mediated gaze awareness. Attention Perception & Psychophysics 62 (04 2012), 586–595. https://doi.org/10.3758/BF03212110

    work page doi:10.3758/bf03212110 2012

  11. [11]

    Matthias Gamer and Heiko Hecht. 2007. Are you looking at me? Measuring the cone of gaze. Journal of Experimental Psychology: Human Perception and Performance 33, 3 (2007), 705–715. https://doi.org/10.1037/0096-1523.33.3.705

    work page doi:10.1037/0096-1523.33.3.705 2007

  12. [12]

    James J Gibson and Anne D Pick. 1963. Perception of another person’s looking behavior. The American journal of psychology 76, 3 (1963), 386–394

    work page 1963

  13. [13]

    VPixx Technologies Inc. 2022. Vpixx. https://https://vpixx.com/

    work page 2022

  14. [14]

    Hessels, Johannes A

    Chiara Jongerius, Roy S. Hessels, Johannes A. Romijn, Ellen M. A. Smets, and Marij A. Hillen. 2020. The Measurement of Eye Contact in Human Interactions: A Scoping Review. Journal of Nonverbal Behavior 44, 3 (01 Sep 2020), 363–389. https://doi.org/10.1007/s10919-020-00333-3

    work page doi:10.1007/s10919-020-00333-3 2020

  15. [15]

    Chris L Kleinke. 1986. Gaze and eye contact: a research review. Psychological bulletin 100, 1 (1986), 78

    work page 1986

  16. [16]

    David J Knight, Daniel Langmeyer, and David C Lundgren. 1973. Eye-contact, distance, and affiliation: The role of observer bias. Sociometry (1973), 390–401

    work page 1973

  17. [17]

    K Krüger and B Hückstedt. 1969. Evaluation of the direction of gazing. Zeitschrift fur Experimentelle und Angewandte Psychologie 16, 3 (1969), 452–472

    work page 1969

  18. [18]

    P K Kuhl, K A Williams, F Lacerda, K N Stevens, and B Lindblom. 1992. Linguistic experience alters phonetic perception in infants by 6 months of age. Science 255, 5044 (Jan. 1992), 606–608

    work page 1992

  19. [19]

    Chris Moore, Philip J Dunham, and Phil Dunham. 2014. Joint attention: Its origins and role in development . Psychology Press

    work page 2014

  20. [20]

    Denis G Pelli and Spatial Vision. 1997. The VideoToolbox software for visual psychophysics: Transforming numbers into movies. Spatial vision 10 (1997), 437–442

    work page 1997

  21. [21]

    Peterson and Miguel P

    Matthew F. Peterson and Miguel P. Eckstein. 2012. Looking just below the eyes is optimal across face recognition tasks. Proceedings of the National Academy of Sciences 109, 48 (2012), E3314–E3323. https://doi.org/10.1073/pnas.1214269109

    work page doi:10.1073/pnas.1214269109 2012

  22. [22]

    Barry Stapley. 1972. Visual enhancement of telephone conversations. (1972)

    work page 1972

  23. [23]

    Lisa J Stephenson, S Gareth Edwards, and Andrew P Bayliss. 2021. From gaze perception to social cognition: The shared-attention system. Perspectives on Psychological Science 16, 3 (2021), 553–576

    work page 2021

  24. [24]

    R Stokes. 1969. Human factors and appearance design considerations of the mod II PICTUREPHONE® station set. IEEE Transactions on Communication Technology 17, 2 (1969), 318–323

    work page 1969

  25. [25]

    Cambridge Research System. 2018. Display++ LCD Monitor. https://www.crsltd.com/

    work page 2018

  26. [26]

    JH White, JR Hegarty, and NA Beasley. 1970. Eye contact and observer bias: A research note. British Journal of Psychology 61, 2 (1970), 271

    work page 1970