Pith. sign in

REVIEW 4 major objections 5 minor 22 references

FIDAC: An Easy-to-use Pipeline to Extract and Interpret Interpersonal Distance From Video

T0 review · 4 major / 5 minor · reviewed 2026-07-31 · grok-4.5

Pith's one-line read FIDAC turns ordinary 2D video into usable interpersonal-distance data by merging facial detectors, optional human coding, and a plane scale.

desk verdict Honest tool announcement for IPD-from-video, but metric claims rest on an unvalidated single-scale benchmark and zero accuracy numbers. read the letter →

arxiv 2607.25146 v1 pith:M4LIBAPB submitted 2026-07-27 cs.CV cs.HC

classification cs.CVcs.HC
keywords nonverbalbehaviorsfacialdetectionvideo-basedtrackingproxemicsinterpersonaldistancehumancoding2Dvideoanalysis
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Interpersonal distance carries a lot of social meaning, but ordinary camera video does not hand it over as clean numbers. This paper presents FIDAC, an open pipeline that takes everyday 2D video, runs multiple facial detectors, fills gaps with a human-coding path when faces are missed, and converts pixel separations into real-world distances with a simple two-point benchmark on the participants’ plane. The aim is to give researchers who study nonverbal behavior a practical way to track who stands where over time without specialized motion-capture hardware. If the method works as described, dynamic proximity and related cues become extractable from smartphone or digital-camera recordings that labs already collect. The authors position the tool as a foundation they will later test across depths and orientations and extend toward broader proxemic measures such as synchrony.

What carries the argument

FIDAC’s three-step pipeline: frame extraction, multi-model facial detection with confidence and size filters plus optional human coding of misses, then merge and plane benchmarking that converts bounding-box pixel separations into real-world units.

What would settle it

Run the full pipeline on videos where true interpersonal distances are independently measured at several depths, angles, and partial occlusions, and check whether the recovered distances stay within a stated error band against those ground-truth values.

Watch

Extended reading notes

Core claim

The paper claims that combining several open-source facial detection models, a structured human-coding fallback for missed faces, and a benchmarking step that scales pixel distances from two known points on the same plane is enough to turn ordinary 2D video into actionable location and interpersonal-distance time series.

Load-bearing premise

A single known real-world distance between two user-picked points on the same plane as the people is enough to turn 2D face-box pixel gaps into accurate interpersonal distances despite unknown depth, pose, and camera geometry.

Editorial extensions

If this is right

  • Labs can extract continuous interpersonal-distance traces from ordinary smartphone or digital-camera recordings instead of only stop-distance snapshots.
  • Missed or blocked faces need not drop entire segments if the human-coding path is used to complete the series.
  • Pixel face locations become interpretable in Hall-style proximity zones once the plane benchmark is applied.
  • The same position series can later feed automated proxemic features such as synchrony once those modules are added.
  • Researchers can swap in other facial detectors without rebuilding the rest of the distance pipeline.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the planned depth-and-orientation benchmark holds, archival interview and meeting video could be reanalyzed for proximity dynamics that were never instrumented at capture time.
  • The human-coding fallback implies a hybrid accuracy model: automated bulk tracking plus sparse expert correction may be more practical for many social-science labs than fully automatic 3D systems.
  • Failure modes at extreme angles or heavy occlusion will likely set the real ceiling on when FIDAC can replace dedicated motion capture rather than complement it.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The manuscript presents FIDAC, an open-source Python library that extracts interpersonal distance (IPD) from ordinary 2D video. The pipeline has three stages: frame extraction via OpenCV, face detection using pre-loaded models (Py-Feat and MediaPipe, with user-swappable alternatives) whose outputs are combined, and a merge step in which model-generated detections are supplemented by a human-coding interface for frames where detection fails. A "benchmarking" tool (§2.3) converts pixel distances to metric units by having the user click two points of known real-world separation on the participants' plane. The paper is a methods description: no accuracy, precision, or error characterization is reported, and the abstract and Conclusion explicitly defer evaluation (varied depths/orientations, comparison against existing detectors, a user study) to future work.

Significance. If the tool performs as described, it addresses a real and well-motivated gap: proxemics researchers currently choose between expensive motion capture, VR telemetry, or tedious manual coding, and an accessible open-source pipeline built on commodity video would lower the barrier substantially. The manuscript ships working software (open-sourced), a sensible human-in-the-loop design for detection failures, and is commendably honest about its own limitations — the authors state plainly that validation at varying depths and orientations is future work rather than overselling the current release. However, the contribution as it stands is a system description without quantitative evidence that the output distances are accurate, and the one component that converts pixels to meters (§2.3) rests on a geometric assumption that is generally violated in realistic filming conditions. The significance is therefore conditional: the engineering is plausible and useful, but the central promise of "actionable data about location and interpersonal distance" is not yet demonstrated.

major comments (4)
  1. [§2.3 Benchmarking, Figure 3] The metric conversion reduces to a single two-point scale ratio (known distance / pixel distance). Under perspective projection, pixels-per-meter is not constant across a plane: it varies with image position and with depth. A single global scale factor is exact only for a fronto-parallel plane at the calibration depth. In the paper's own motivating example (two participants seated across a table filmed by a smartphone), an oblique camera angle makes the near and far ends of the table subtend different pixels-per-meter, so the same physical distance yields different pixel distances depending on where in the frame the participants sit. The authors should either (a) implement the standard minimal fix — a four-point planar homography on the ground/table plane, which is cheap and available in OpenCV — or (b) state the fronto-parallel assumption explicitly and provide an error bound or worked
  2. [§2.3 / §2.2 interaction] The quantity measured is the separation of face bounding boxes, which sit above the calibrated plane (e.g., above the tabletop or floor). Unless the camera is at face height and faces remain in a single plane parallel to the image plane, projecting face positions onto the calibrated plane introduces an additional, unmodeled error. This matters most exactly where the signal lives: the IPD dynamics of interest (leaning in, approaching) change the subject's depth relative to the camera, moving them off the calibration depth and inflating or deflating apparent distance. The manuscript never discusses this geometry. At minimum the paper needs a stated model of how 2D face-box positions are interpreted as participant positions (box centers? feet estimated from box bottom?) and a discussion of when the approximation breaks down.
  3. [§2.2 Analysis + Facial Detection] Two load-bearing algorithmic details are unspecified. (1) Identity association: IPD requires knowing which bounding box belongs to which participant in every frame, but no tracking or data-association method is described — the text only says faces are detected per frame. For two or more people who move, cross, or temporarily occlude each other, naive per-frame detection will swap identities and corrupt the distance time series. (2) Multi-model fusion: the text says models are combined so that 'discrepancies in one are corrected by another,' but no fusion rule is given (union? intersection? confidence-weighted vote? temporal gap-filling?). Since the multi-model merge is one of the paper's three headline contributions, the merge logic needs to be specified precisely enough to reproduce.
  4. [Whole manuscript; Abstract and §3] No quantitative evaluation of any kind is reported — no detection accuracy on a benchmark clip, no distance error versus ground truth at any depth or orientation, not even a worked example on a staged video with a tape-measure reference. The authors are transparent that evaluation is future work, which is appropriate for a demo or workshop track, but for a journal publication the central claim ('transforms facial detection results into actionable data about location and interpersonal distance') requires at least minimal evidence. A low-cost experiment is within the paper's scope: film two people at known separations (e.g., 0.5–4 m) at a few camera angles, and report measured-vs-true distance error. This would also directly address the §2.3 concerns above.
minor comments (5)
  1. [§1] Hall's zone boundaries are given as exact centimeter cutoffs (e.g., intimate 0–46 cm); since FIDAC's output will be compared against these zones, a sentence acknowledging that zone boundaries are approximate cultural averages would help readers interpret threshold-adjacent measurements.
  2. [§2.1] The phrase 'FIDAC's algorithms were trained to identify faces' is misleading — FIDAC wraps pre-trained detectors (Py-Feat, MediaPipe); it does not train them. Please reword to avoid implying the authors trained the detection models.
  3. [§2.2] Guidance on the free parameters (confidence threshold, min/max face size, frames-per-second) is qualitative only. Default values used by the authors, and a note on how frame subsampling interacts with the temporal resolution needed for IPD dynamics, would aid reproducibility.
  4. [§3] 'Unitary values' should presumably read 'real-world units' or 'metric values.'
  5. [General] The manuscript states the library is open-sourced but no repository URL, license, or version appears in the text provided; please add a link (anonymized if required). An OpenCV citation is also missing. Figures 1–3 are described but their resolution/legibility should be checked at print size.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: FIDAC is a descriptive software pipeline, not a first-principles or predictive claim chain.

full rationale

The manuscript presents an engineering workflow (frame extraction → multi-model face detection → optional human coding → two-point plane scale benchmarking → CSV merge) built on external detectors (Py-Feat, MediaPipe) and OpenCV. It does not claim to derive interpersonal-distance laws, fit parameters and re-predict them, or import uniqueness theorems from the authors’ prior work. Self-citations (e.g., Bailenson 2003; Han et al. 2023) appear only as background on proxemics/VR, not as load-bearing justifications that force the pipeline’s outputs. The benchmarking step is a user-supplied scale ratio, not a fitted-then-predicted quantity. Effectiveness evaluation is explicitly deferred to future work. Under the circularity criteria there is therefore nothing that reduces by construction to its own inputs; score 0 with empty steps is the correct outcome. Correctness concerns about single-ratio depth handling are orthogonal to circularity.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The central claim rests on engineering assumptions about 2D vision and a simple homography-free scale, not on fitted physical constants. Load-bearing premises are that off-the-shelf face detectors plus optional human labels yield usable trajectories, and that a single in-plane reference distance converts pixel face separations into real-world IPD with acceptable depth error. No new physical entities are postulated; free parameters are operational thresholds chosen by the user.

free parameters (4)
  • face confidence threshold = not specified (user-chosen; high threshold recommended)
    User-set cutoff for accepting detector outputs; directly controls false positives/negatives and thus which frames contribute to IPD.
  • face size limits (min/max bounding-box dimensions)
    Ad hoc filters to reject false-positive detections whose dimensions fall outside expected face scale.
  • frames analyzed per second = example values 1 or 60 fps mentioned; default unspecified
    Subsampling rate chosen for speed vs temporal resolution; affects trajectory smoothness and measured distance dynamics.
  • benchmark plane reference distance = scene-dependent (user-supplied)
    Known real-world length between two user-clicked points used as the sole scale factor from pixels to physical units.
assumptions (4)
  • domain assumption Off-the-shelf 2D face detectors (MediaPipe, Py-Feat, or user-supplied) produce bounding boxes whose centers/edges are adequate proxies for person location for IPD.
    Invoked throughout §2.2; no error model relating box geometry to body centroid is given.
  • ad hoc to paper A single known distance between two points on the same plane as participants is sufficient to convert pixel separations into real-world interpersonal distances despite perspective and depth variation.
    Core of the benchmarking tool in §2.3; treats scale as uniform on that plane and does not model out-of-plane motion.
  • domain assumption Human coding can reliably fill detector misses without introducing systematic bias larger than the scientific effect of interest.
    Stated as motivation and pipeline step in Introduction and §2; no inter-rater protocol or reliability metric is specified.
  • domain assumption Stable camera pose (e.g., tripod) keeps the pixel-to-world mapping valid across the video.
    Noted in §2.1 Video Setup as a requirement for relative-position computation.
invented entities (1)
  • FIDAC pipeline (multi-model merge + human-coding file + plane benchmark as an integrated IPD library)
    purpose: Package existing detectors and a simple scale step into a researcher-facing tool for dynamic interpersonal distance from 2D video.
    The named system is the paper's contribution; it is software architecture rather than a new physical or statistical entity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FIDAC: An Easy-to-use Pipeline to Extract and Interpret Interpersonal Distance From Video." pith.science (2026). https://pith.science/paper/M4LIBAPB

@misc{pith2026260725146,
  author       = {Pith},
  title        = {Pith review of: FIDAC: An Easy-to-use Pipeline to Extract and Interpret Interpersonal Distance From Video},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/M4LIBAPB}},
  note         = {Machine review of arXiv:2607.25146}
}
read the original abstract

The distance between persons reveals significant information about their perception of each other. However, such information is not easily extractable and interpretable from video input. We developed an open-sourced library, Facial Interpersonal Distance Analysis and Coding (FIDAC) that transforms facial detection results into actionable data about location and interpersonal distance. This tool merges data from multiple open-source facial detection models, strategically compensating for gaps in any individual model. In addition, we include methods for more accurate tracking, such as a pipeline for human coding of the selection of faces and a benchmarking tool to reduce depth distortion. For next steps, we plan on building upon FIDAC by evaluating its effectiveness at measuring interpersonal distance at various depths and orientations while further integrating features of proxemic analysis such as synchrony into its software.

Figures

Figures reproduced from arXiv: 2607.25146 by the authors.

Figure 2
Figure 2. A detected face represented as a bounding box [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 1
Figure 1. FIDAC’s three-step pipeline for analyzing inter [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Benchmarking tool used to convert pixelated dis [PITH_FULL_IMAGE:figures/full_fig_p002_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

22 extracted references · 1 canonical work pages

  1. [1]

    Bailenson, Jim Blascovich, Andrew C

    Jeremy N. Bailenson, Jim Blascovich, Andrew C. Beall, and Jack M. Loomis. 2003. Interpersonal Distance in Immersive Virtual Environments.Personality and Social Psychology Bulletin29, 7 (2003), 819–833

  2. [2]

    Jonathan Baumann, Björn Krüger, Alexander Zinke, and Andreas Weber. 2011. Data-Driven Completion of Motion Capture Data. InProceedings of the Euro- graphics/ACM SIGGRAPH Workshop on Posters and Demos

  3. [3]

    1993.How Humans Relate: A New Interpersonal Theory

    John Birtchnell. 1993.How Humans Relate: A New Interpersonal Theory. Praeger, Westport, CT

  4. [4]

    Ronnie Bryan, Pietro Perona, and Ralph Adolphs. 2012. Perspective Distortion from Interpersonal Distance Is an Implicit Visual Cue for Social Judgments of Faces.PLOS ONE7, 9 (2012), e45301

  5. [5]

    Jin Hyun Cheong, Tiankang Xie, Sophie Byrne, and Luke J. Chang. 2021. Py-Feat: Python Facial Expression Analysis Toolbox.arXiv preprint arXiv:2104.03509 (2021)

  6. [6]

    Kishor Das, Thiago de Paula Oliveira, and John Newell. 2023. Comparison of Markerless and Marker-Based Motion Capture Systems Using 95% Functional Limits of Agreement in a Linear Mixed-Effects Modelling Framework.Scientific Reports13 (2023), 22880

  7. [7]

    Edward T. Hall. 1966.The Hidden Dimension. Doubleday, Garden City, NY

  8. [8]

    Miller, Cyan DeVeaux, Hanseul Jun, Kristine L

    Eugy Han, Mark R. Miller, Cyan DeVeaux, Hanseul Jun, Kristine L. Nowak, Jef- frey T. Hancock, Nilam Ram, and Jeremy N. Bailenson. 2023. People, Places, and Time: A Large-Scale, Longitudinal Study of Transformed Avatars and Environ- mental Context in Group Interaction in the Metaverse.Journal of Computer- Mediated Communication28, 2 (2023), zmac031

Show all 22 references
  1. [9]

    Leslie A. Hayduk. 1978. Personal Space: An Evaluative and Orienting Overview. Psychological Bulletin85, 1 (1978), 117–134

  2. [10]

    Leslie A. Hayduk. 1983. Personal Space: Where We Now Stand.Psychological Bulletin94, 2 (1983), 293–335

  3. [11]

    Sigurðsson, Kayla D

    Naoaki Ito, Haraldur B. Sigurðsson, Kayla D. Seymore, Elanna K. Arhos, Thomas S. Buchanan, Lynn Snyder-Mackler, and Karin Grävare Silbernagel. 2022. Marker- less Motion Capture: What Clinician-Scientists Need to Know Right Now.JSAMS Plus1 (2022), 100001

  4. [12]

    Distance

    Sidney M. Jourard and Robert Friedman. 1970. Experimenter-Subject "Distance" and Self-Disclosure.Journal of Personality and Social Psychology15, 3 (1970), 278–282

  5. [13]

    Dimitrios Kollias and Stefanos Zafeiriou. 2018. Aff-Wild2: Extending the Aff-Wild Database for Affect Recognition. arXiv:1811.07770

  6. [14]

    Leon O. H. Kroczek, Michael Pfaller, Bastian Lange, Mathias Müller, and Andreas Mühlberger. 2020. Interpersonal Distance during Real-Time Social Interaction: Insights from Subjective Experience, Behavior, and Physiology.Frontiers in Psychiatry11 (2020), 561

  7. [15]

    Donna M. Lloyd. 2009. The Space between Us: A Neurophilosophical Framework for the Investigation of Human Interpersonal Space.Neuroscience & Biobehavioral Reviews33, 3 (2009), 297–304

  8. [16]

    Camillo Lugaresi, Jiuqiang Tang, Hadon Nash, Chris McClanahan, Esha Uboweja, Michael Hays, Fan Zhang, Chuo-Ling Chang, Ming Guang Yong, Juhyun Lee, Wan-Teh Chang, Wei Hua, Manfred Georg, and Matthias Grundmann. 2019. MediaPipe: A Framework for Building Perception Pipelines. ar...

  9. [17]

    Shushi Namba, Wataru Sato, Masaki Osumi, and Koh Shimokawa. 2021. Assessing Automated Facial Action Unit Detection Systems for Analyzing Cross-Domain Facial Expression Databases.Sensors21, 12 (2021), 4222

  10. [18]

    Shamay-Tsoory

    Anat Perry, Orly Rubinsten, Leehe Peled, and Simone G. Shamay-Tsoory. 2013. Don’t Stand So Close to Me: A Behavioral and ERP Study of Preferred Interper- sonal Distance.NeuroImage83 (2013), 761–769

  11. [19]

    Virginie Quintard, Stéphane Jouffre, Jean-Claude Croizet, and Cedric A. Bouquet

  12. [20]

    Veronica Romero, Joseph Amaral, Paula Fitzpatrick, R. C. Schmidt, Ashley W. Duncan, and Michael J. Richardson. 2017. Can Low-Cost Motion-Tracking Sys- tems Substitute a Polhemus System when Researching Social Motor Coordination in Children?Behavior Research Methods49, 2 (2017)...

  13. [21]

    Rosenberger, Michael Naef, Christoph Eisenegger, and Claus Lamm

    Lisa A. Rosenberger, Michael Naef, Christoph Eisenegger, and Claus Lamm. 2020. Interpersonal Distance Adjustments after Interactions with a Generous and Selfish Trustee during a Repeated Trust Game.Journal of Experimental Social Psychology90 (2020), 104001

  14. [2020]

    doi:10.1007/s00426-018-0981-z

    The Influence of Passionate Love on Self-Other Discrimination during Joint Action.Psychological Research84, 1 (2020), 51–61. doi:10.1007/s00426-018-0981-z

Pith tools

Reviewed July 31, 2026 · model on record in the stance chip above.