Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Lights, Camera, Matching: The Role of Image Illumination in Fair Face Recognition

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper argues that controlling the illumination of face images at acquisition time can shrink the demographic accuracy gap in face recognition, because pairing images with similar brightness reduces the distance d' between the…

desk verdict Useful descriptive study of brightness-based subset selection in face matching; the causal acquisition-time claim is unsupported by the missing control subsets. read the letter →

arxiv 2501.08910 v1 pith:DURHJNGR submitted 2025-01-15 cs.CV

Gabriella Pangelinan , Grace Bezold , Haiyu Wu , Michael C. King , Kevin W. Bowyer This is my paper · ORCID
classification cs.CV
keywords facerecognitiondemographicbiasilluminationbrightnessbalancingmatedpairsd-primeimagequalityMORPH
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

Face recognition systems compare two photos of the same person and produce a similarity score; for the same matcher, those score distributions sit differently for Caucasian (CF) and African American (AF) women. This paper tests whether balancing the illumination of the face region across the two groups narrows that gap, measured as d' between the genuine-match score distributions. Three experiments on the MORPH mugshot dataset show that restricting to pairs whose face-skin brightness is similar (median-pixel difference ≤ 0.5) reduces the d' gap by 46.8%, and restricting to pairs whose brightness distributions are bi- or multimodal reduces it by 57.6%, while also raising mean scores for both groups. If correct, the result says that how images are lit at capture time — not just algorithm design — is a controllable lever for reducing demographic bias in face recognition.

What carries the argument

Three balancing factors carry the argument. (1) Brightness value (BV): the median grayscale pixel value of the face skin region extracted by face parsing; brightness value difference (BVD) is the absolute difference between the two images' BVs in a mated pair, and low BVD means the two photos are lit similarly. (2) Brightness distribution modality (BDM): the pixel-value histogram of the face skin region is labeled unimodal, bimodal, or multimodal via smoothed-peak detection (smoothing window 4, relative threshold 0.5); unimodality signals poor illumination, predominantly overexposure in CF images and broader low-brightness peaks in AF images. (3) Brightness-distribution intersection-over-union (BD-IoU): for a set containing a CF pair and an AF pair, the overlap between the four brightness distributions is computed under the two possible image matchings and the maximum average is taken, quantifying how similarly illuminated the two pairs are. The outcome metric throughout is d' — the separation between the CF and AF genuine-score distributions in standard-deviation units — and the paper reports the percent shift in d' ('d' shift') relative to the baseline.

What would settle it

Select from the original MORPH pools the same numbers of CF and AF pairs at random, or matched on a different quality factor such as sharpness, and compute the d' shift for those subsets; if random or non-illumination quality-matched subsets show d' reductions comparable to the 30–58% reported, the brightness-balancing explanation is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that the CF-AF accuracy gap in mated face-image matching is driven in part by how similarly and how well the two images in a pair are illuminated, and that balancing illumination across demographics shrinks the gap while improving accuracy. Using a curated subset of MORPH (Caucasian and African American female images) and an ArcFace-based matcher, the authors compute d' between the distributions of genuine similarity scores for the two groups. They report a 46.8% decrease in d' when both groups are restricted to the 1,000 mated pairs with the smallest brightness value difference (mean BVD 0.5), a 57.6% decrease when pairs with unimodal brightness distributions are excluded (keeping only bi-/multimodal pairs), and a 30.6% decrease when pairs are selected so that the brightness distributions of a CF pair share high intersection-over-union with those of an AF pair. In each balanced subset the mean genuine score rises for both demographics, with CF improving more than AF, which narrows the gap from both sides.

Load-bearing premise

The cause of the gap reduction is the brightness-balancing property itself, not the general effect of choosing a smaller, more homogeneous, higher-quality subset of mated pairs for both demographic groups.

Editorial extensions

If this is right

  • If illumination is controllable at capture time, ID-photo pipelines (driver's licenses, passports, mugshots) could adopt brightness-matching requirements for paired images and reduce demographic accuracy gaps without retraining the matcher.
  • Brightness value difference is a simple, interpretable quality metric: keeping mated pairs with BVD ≤ 0.5 (and ideally below 1) is associated with a 30–47% smaller gap.
  • Detecting unimodal face-skin brightness histograms can flag poorly illuminated images; because unimodality for CF images is largely overexposure (pixels near 240–255), such a flag could prompt re-acquisition.
  • Because mean scores improve for both demographics in every balanced subset, illumination balancing appears to reduce false non-match rate as well as the group gap.

Reading between the lines

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

  • The paper demonstrates a selection effect, not yet a causal capture-time intervention; a direct test would be to re-photograph subjects under varied lighting and verify that the d' reduction survives when illumination is manipulated rather than selected.
  • Without a random-subset control of equal size, some of the d' reduction may be a generic 'smaller, cleaner subset' effect; selecting same-size random subsets or subsets matched on another quality factor (e.g., sharpness) would isolate the illumination-specific contribution.
  • The BDM result suggests an asymmetry: overexposure is the dominant failure mode for CF images, while AF unimodal images can be over- or underexposed, so a single global 'well-illuminated' criterion may be less fair than demographics-aware exposure targets.
  • The balancing factors tested here rely on histograms of the face skin region; an extension would test whether the same reductions hold in unconstrained 'in-the-wild' collections, where ambient illumination varies far more than in mugshot-style images.
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

3 major / 5 minor

Summary. The paper studies whether balancing the brightness of face-skin regions between Caucasian female (CF) and African American female (AF) mated image pairs reduces the demographic gap in face-recognition similarity-score distributions, measured by d'. Three subset-selection experiments are reported: Brightness Value Difference (BVD), where pairs are ordered by the absolute difference of median brightness values; Brightness Distribution Modality (BDM), where pairs are grouped by whether their per-image brightness histograms are unimodal, bimodal, or multimodal; and Brightness Distribution IoU (BD-IoU), where CF and AF pairs are matched by intersection-over-union of their brightness distributions. In each experiment, d' is recomputed on the selected subset of CF/AF pairs, yielding reported decreases of up to 46.8%, 57.6%, and 30.6% respectively, together with small increases in mean similarity scores. The authors conclude that illumination should be controlled at acquisition time to reduce demographic accuracy differences.

Significance. If the reported reductions are causally attributable to illumination balancing, the result would be practically valuable for operational face recognition under controlled capture, and the paper would add a useful pair-level perspective to the existing single-image brightness-quality literature. The work has several strengths: it uses a standard matcher (ArcFace/Glint360k), a controlled mugshot dataset, transparent performance metrics, three complementary operationalizations of brightness, and detailed supplementary tables that make the descriptive computations easy to follow. The paper does not provide code, but the algorithms and tables are sufficiently specified to reproduce the main numerical claims. The central weakness is that all three experiments are observational subset-selection studies: d' is measured on pre-existing pairs chosen by brightness criteria, with no random-subset control, no variance estimate, and no intervention on acquisition conditions. As a result, the descriptive finding that brightness-based subsets have reduced d' is probably sound, but the causal acquisition-time conclusion in Sec. 8 is not yet supported.

major comments (3)
  1. [Secs. 4.2, 5.2, 6.2; Sec. 8] The central causal claim of Sec. 8—that acquisition-time control of illumination is 'essential' for reducing demographic gaps—is not supported by the present design, because every experiment estimates d' on subsets selected by the brightness factor itself and no control condition is reported. A random subset of 1,000 CF and 1,000 AF pairs of the same size could plausibly show a large d' reduction merely by excluding the tails of the score distributions, and the BVD result in Sec. 4.2 (d' shift -46.8% for the top 1k pairs, mean BVD 0.5) is never compared with such a null. I request a bootstrap or permutation null over random subsets of each N, plus a control subset matched on an orthogonal quality factor such as blur, resolution, or sharpness. Without these, the observed d' declines are compatible with a generic homogeneity/quality-selection effect rather than with the brightness-balancing mechanism.
  2. [Sec. 5.2, Tab. 3; Sec. 6.2, Tab. 5] The BDM and BD-IoU experiments simultaneously vary more than brightness balancing. In Sec. 5.2, all balanced subsets exclude pairs containing a Uni image, a category the paper itself associates with overexposure and worse scores for CF; dropping these pairs may reduce d' even if the property 'the two images are similarly illuminated' plays no causal role. In Sec. 6.2, the BD-IoU construction both excludes Uni images and requires high distribution overlap across a CF and an AF pair, so the contribution of each requirement cannot be separated. The narrative in Sec. 7.2 that unimodality 'indicates poorly illuminated images' is inferred from the same data used to define the subsets, not from an independent quality label. I recommend an ablation that keeps the exclusion of non-Uni images fixed while varying only the brightness-difference/overlap criterion, and a comparison with subsets balanced on a non-brightness quality factor.
  3. [Secs. 3, 4.2, 5.2; Tabs. 1–3; Supplementary Material] No uncertainty estimates are given for the headline d' shifts. The BVD and BD-IoU tables report single values for deterministic top-N subsets; the BDM experiment averages over 10 shuffles but reports no variance or confidence interval. It is therefore impossible to judge whether -46.8%, -57.6%, or -30.6% are distinguishable from the sampling variability of any 1k-5k subset. In addition, the matching procedure in Sec. 4.2 reports '33,735 total matched pairs' although there are 33,470 CF mated pairs; the pool size and matching rule should be stated precisely, since the number of available unique matches constrains what the top-N subsets represent.
minor comments (5)
  1. [Sec. 5.3] The word 'sigfinicantly' should be 'significantly'.
  2. [Sec. 5.1, Tab. 2] The text states that UniBi pairs have a d' increase of 2.5%, while Table 2 reports 2.8%; please reconcile the values.
  3. [Sec. 5.3] Figure 8 is referenced before Figure 7 in the prose; reorder the figures or their citations.
  4. [Supplementary Material, BDM] The modality parameters SW=4 and RT=0.5 were selected after manual inspection; please report how sensitive Tables 2 and 3 are to these parameter choices, or at least state that the main conclusions are stable across a small neighborhood of (SW, RT).
  5. [Sec. 7.1, Tab. 7] The statement 'the brightness difference is ≤ 0.5 (as it was when taking the 1k Top Pairs)' is imprecise: Table 7 reports the mean BVD for that subset as 0.486 with a standard deviation of 0.5, so many included pairs have BVD greater than 0.5. The wording should say 'mean BVD' rather than implying a hard threshold on every pair.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the reported d' reductions are measurements on brightness-selected subsets, not predictions from fitted parameters; the acquisition-time causal claim is an external-validity concern, not a circular one.

full rationale

The paper's chain is: define brightness-based balancing factors (BVD, BDM, BD-IoU), select subsets of existing CF/AF mated pairs that are balanced on those factors, and then measure the mated-score d' on the selected subsets. None of the d' values is derived from an equation involving the balancing factor: the balancing factors are computed from face-region pixel brightness, while d' is computed from matcher cosine similarities. The reported 'd' shift' is a conditional summary of the selected data, not a prediction from a fitted parameter. The modality thresholds (SW=4, RT=0.5) are chosen manually and the subset sizes N are arbitrary, which weakens the generalizability of the headline percentages, but this is researcher discretion rather than circularity because the outcome is not defined in terms of the selection rule. The only self-citations ([5], [23]) supply the dataset and prior related-work framing; the central claim does not rest on an unverified self-cited theorem or uniqueness result. The Sec. 8 conclusion that acquisition-time illumination control is 'essential' goes beyond the observational subset-selection evidence, but an unsupported causal inference is a validity problem, not a circular derivation. No step reduces by construction to its own input, so the paper receives a circularity score of 0.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The central claims rest on three chosen thresholds or subset sizes, plus domain assumptions that MORPH represents operational capture, that the skin-segmentation mask is demographic-neutral, and that a single ArcFace matcher's scores are the right outcome measure. No new physical or theoretical entities are posited.

free parameters (3)
  • BDM smoothing window SW = 4
    Chosen by manual inspection of smoothed histograms in the Supplementary Material; affects peak detection and thus all BDM balancing results (Tab. 3).
  • BDM relative threshold RT = 0.5
    Chosen as the minimum peak height as a fraction of the maximum count; hand-tuned, affects modality labels and d' shifts.
  • Balanced subset size N = 1k, 3k, 5k, 10k (varies by experiment)
    Subset sizes are chosen by the authors; d' shift magnitudes depend strongly on N (for BVD, the shift goes from -24.2% at 10k to -46.8% at 1k). No justification is given for the chosen values.
assumptions (3)
  • domain assumption MORPH images captured under controlled illumination are a valid proxy for operational photo capture scenarios such as mugshots, driver's licenses, and passports.
    Invoked in Sec. 8 to generalize the subset results to acquisition-time interventions.
  • domain assumption The BiSeNet face-skin mask provides a consistent and unbiased segmentation of the face skin region across Caucasian and African American faces.
    All brightness measurements depend on this mask, and no demographic-specific validation is reported (Sec. 3.3).
  • domain assumption The ArcFace-based matcher's cosine similarity scores are comparable and meaningful across demographic groups.
    The entire analysis uses this matcher's score distributions as the outcome variable, with no cross-matcher validation (Sec. 3.1).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Lights, Camera, Matching: The Role of Image Illumination in Fair Face Recognition." pith.science (2026). https://pith.science/paper/DURHJNGR

@misc{pith2026250108910,
  author       = {Pith},
  title        = {Pith review of: Lights, Camera, Matching: The Role of Image Illumination in Fair Face Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DURHJNGR}},
  note         = {Machine review of arXiv:2501.08910}
}
read the original abstract

Facial brightness is a key image quality factor impacting face recognition accuracy differentials across demographic groups. In this work, we aim to decrease the accuracy gap between the similarity score distributions for Caucasian and African American female mated image pairs, as measured by d' between distributions. To balance brightness across demographic groups, we conduct three experiments, interpreting brightness in the face skin region either as median pixel value or as the distribution of pixel values. Balancing based on median brightness alone yields up to a 46.8% decrease in d', while balancing based on brightness distribution yields up to a 57.6% decrease. In all three cases, the similarity scores of the individual distributions improve, with mean scores maximally improving 5.9% for Caucasian females and 3.7% for African American females.

Figures

Figures reproduced from arXiv: 2501.08910 by the authors.

Figure 2
Figure 2. Data pre-processing pipeline. 4. Brightness Value Difference 4.1. Pre-Processing We begin by calculating the brightness value (BV) for each image as the median brightness of the pixels in the face skin region. We use the median rather than mean, given that some of the brightness value distributions exhibit significant skew. (This idea will be explored in-depth in Sec. 5.) Example CF / AF images with the minimum, ave… view at source ↗
Figure 3
Figure 3. Ex. images with min., avg., and max. BVs. [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Example pairs with the min, avg., and max BVDs. [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (6 more)
Figure 5
Figure 5. Figure 5: Relationship between mean BVD (x¯b) decrease and d’ decrease. 5. Brightness Distribution Modality 5.1. Pre-Processing Each image’s distribution of brightness pixels can be described as: unimodal (“Uni”), bimodal (“Bi”), or multimodal (“Multi”). (Our method for assign￾i…
Figure 6
Figure 6. Figure 6: Example Uni / Bi / Multi images (and corresponding distributions) for the same CF subject. [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: shows example CF images with increasing percentages of their face skin pixels > 240. As this percentage increases, the face appears more overex￾posed - and facial feature information is lost / ob￾scured. A total of 2197 CF Uni images (40%) are at least as overexposed a…
Figure 8
Figure 8. Figure 8: For Uni images, overlaid brightness distributions (left) and heatmaps highlighting the fraction of pixels with BVs [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: BV and BVD distributions. The full results of the BVD experiment are shown in Tab. 7. For each balanced N Top Pairs subset, we report the mean similarity score (“Score x¯b”), shift in x¯b vs. the baseline mean score (“Score x¯b” Shift), d’ between CF-AF, d’ shift vs. t…
Figure 10
Figure 10. Figure 10: Relative frequency distributions of pixel values for Uni, Bi, and Multi images. [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Review of Demographic Fairness in Face Recognition

    cs.CV 2025-02 conditional novelty 3.0 of 10

    A structured review of demographic fairness in face recognition covering causes, datasets, assessment metrics, and mitigation methods.

Reference graph

Works this paper leans on

26 extracted references · 25 canonical work pages · cited by 1 Pith paper

  1. [1]

    https : / / github

    Insightface: 2d and 3d face analysis project. https : / / github . com / deepinsight / insightface/tree/master/model_zoo. 2

    work page

  2. [2]

    Design and evaluation of photometric image quality measures for effec- tive face recognition

    Ayman Abaza, Mary Ann Harrison, Thirimachos Bourlai, and Arun Ross. Design and evaluation of photometric image quality measures for effec- tive face recognition. IET Biometrics, 3(4):314–324,

    work page

  3. [3]

    Review on the effects of age, gender, and race demographics on automatic face recognition

    Salem Hamed Abdurrahim, Salina Abdul Samad, and Aqilah Baseri Huddin. Review on the effects of age, gender, and race demographics on automatic face recognition. The Visual Computer , 34:1617–1630,

    work page

  4. [4]

    How does gender balance in training data affect face recognition accuracy? In 2020 IEEE International Joint Conference on Biometrics (IJCB) , pages 1–10

    V ´ıtor Albiero, Kai Zhang, and Kevin W Bowyer. How does gender balance in training data affect face recognition accuracy? In 2020 IEEE International Joint Conference on Biometrics (IJCB) , pages 1–10. IEEE, 2020. 1

    work page 2020

  5. [5]

    Gendered differences in face recognition accuracy explained by hairstyles, makeup, and facial morphology

    V ´ıtor Albiero, Kai Zhang, Michael C King, and Kevin W Bowyer. Gendered differences in face recognition accuracy explained by hairstyles, makeup, and facial morphology. IEEE Transactions on Information Forensics and Security, 17:127–137,

    work page

  6. [6]

    Partial fc: Training 10 million identities on a single machine

    Xiang An, Xuhan Zhu, Yuan Gao, Yang Xiao, Yongle Zhao, Ziyong Feng, Lan Wu, Bin Qin, Ming Zhang, Debing Zhang, et al. Partial fc: Training 10 million identities on a single machine. In ICCV, pages 1445– 1449, 2021. 2

    work page 2021

  7. [7]

    The gender gap in face recogni- tion accuracy is a hairy problem

    Aman Bhatta, V ´ıtor Albiero, Kevin W Bowyer, and Michael C King. The gender gap in face recogni- tion accuracy is a hairy problem. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 303–312, 2023. 1

    work page 2023

  8. [8]

    Accuracy comparison across face recognition algorithms: Where are we on measuring race bias? IEEE Transactions on Biomet- rics, Behavior, and Identity Science , 3(1):101–111,

    Jacqueline G Cavazos, P Jonathon Phillips, Carlos D Castillo, and Alice J O’Toole. Accuracy comparison across face recognition algorithms: Where are we on measuring race bias? IEEE Transactions on Biomet- rics, Behavior, and Identity Science , 3(1):101–111,

    work page

Show all 26 references
  1. [9]

    Arcface: Additive angular margin loss for deep face recognition

    Jiankang Deng, Jia Guo, Niannan Xue, and Stefanos Zafeiriou. Arcface: Additive angular margin loss for deep face recognition. In CVPR, 2019. 2

    2019

  2. [10]

    De- mographic bias in biometrics: A survey on an emerg- ing challenge

    Pawel Drozdowski, Christian Rathgeb, Antitza Dantcheva, Naser Damer, and Christoph Busch. De- mographic bias in biometrics: A survey on an emerg- ing challenge. IEEE Transactions on Technology and Society, 1(2):89–103, 2020. 1

    2020

  3. [11]

    Miti- gating face recognition bias via group adaptive clas- sifier

    Sixue Gong, Xiaoming Liu, and Anil K Jain. Miti- gating face recognition bias via group adaptive clas- sifier. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 3414–3424, 2021. 1

    2021

  4. [12]

    Face recognition vendor test (FRVT) part 8: Summarizing demographic differentials , vol- ume 8429

    Patrick Grother. Face recognition vendor test (FRVT) part 8: Summarizing demographic differentials , vol- ume 8429. National Institute of Standards and Tech- nology, 2022. 2

    2022

  5. [13]

    Face recognition vendor test (FRVT): Part 3, demo- graphic effects

    Patrick Grother, Mei Ngan, and Kayee Hanaoka. Face recognition vendor test (FRVT): Part 3, demo- graphic effects. National Institute of Standards and Technology, 2019. 1

    2019

  6. [14]

    Face recognition performance: Role of demographic infor- mation

    Brendan F Klare, Mark J Burge, Joshua C Klontz, Richard W V order Bruegge, and Anil K Jain. Face recognition performance: Role of demographic infor- mation. IEEE Transactions on Information Forensics and Security, 7(6):1789–1801, 2012. 1

    2012

  7. [15]

    Face recognition vendor test (FRVT) performance of auto- mated gender classification algorithms

    Mei Ngan, Patrick J Grother, and Mei Ngan. Face recognition vendor test (FRVT) performance of auto- mated gender classification algorithms . US Depart- ment of Commerce, National Institute of Standards and Technology, 2015. 1

    2015

  8. [16]

    Kagan Ozturk, Haiyu Wu, and Kevin W Bowyer. Can the accuracy bias by facial hairstyle be reduced through balancing the training data? In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1519–1528, 2024. 1

    2024

  9. [17]

    Morph: A longitudinal image database of normal adult age- progression

    Karl Ricanek and Tamirat Tesafaye. Morph: A longitudinal image database of normal adult age- progression. In 7th International Conference on Automatic Face and Gesture Recognition (FGR06) , pages 341–345. IEEE, 2006. 2

    2006

  10. [18]

    Retinaface: Deep face detection model

    Sefik Ilkin Serengil. Retinaface: Deep face detection model. https://github.com/serengil/ retinaface, 2024. Accessed: 2024-11-26. 3

    2024

  11. [19]

    Raja, and Arjan Kuijper

    Philipp Terh ¨orst, Marco Huber, Naser Damer, Flo- rian Kirchbuchner, Kiran B. Raja, and Arjan Kuijper. Pixel-level face image quality assessment for explain- able face recognition. CoRR, abs/2110.11001, 2021. 2

    2021 arXiv

  12. [20]

    Post- comparison mitigation of demographic bias in face recognition using fair score normalization

    Philipp Terh ¨orst, Jan Niklas Kolf, Naser Damer, Florian Kirchbuchner, and Arjan Kuijper. Post- comparison mitigation of demographic bias in face recognition using fair score normalization. Pattern Recognition Letters, 140:332–338, 2020. 1

    2020

  13. [21]

    Comparison-level mitigation of ethnic bias in face recognition

    Philipp Terh ¨orst, Mai Ly Tran, Naser Damer, Florian Kirchbuchner, and Arjan Kuijper. Comparison-level mitigation of ethnic bias in face recognition. In 2020 8th International Workshop on Biometrics and Foren- sics (IWBF), pages 1–6. IEEE, 2020. 1

    2020

  14. [22]

    Characterizing the variability in face recognition accuracy relative to race

    Kushal Vangara, Michael C King, Vitor Albiero, Kevin Bowyer, et al. Characterizing the variability in face recognition accuracy relative to race. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pages 0– 0, 2019. 1

    2019

  15. [23]

    Face recognition accuracy across demographics: Shining a light into the problem

    Haiyu Wu, V ´ıtor Albiero, KS Krishnapriya, Michael C King, and Kevin W Bowyer. Face recognition accuracy across demographics: Shining a light into the problem. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1041–1050, 2023. 2

    2023

  16. [24]

    balanced

    Haiyu Wu and Kevin W Bowyer. What should be balanced in a “balanced”’ face recognition dataset? arXiv preprint arXiv:2304.09818, 2023. 1

    2023 arXiv

  17. [25]

    Fa- cial hair area in face recognition across demograph- ics: Small size, big effect

    Haiyu Wu, Sicong Tian, Aman Bhatta, Ka ˘gan ¨Ozt¨urk, Karl Ricanek, and Kevin W Bowyer. Fa- cial hair area in face recognition across demograph- ics: Small size, big effect. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 1131–1140, 2024. 1

    2024

  18. [26]

    Score ¯xb

    Yakhyo. Face parsing: Facial region segmenta- tion. https://github.com/yakhyo/face- parsing, 2024. Accessed: 2024-11-26. 3 Supplementary Material Brightness Value Difference (BVD) Each image is assigned a brightness value (BV) based on the median pixel value of the face skin r...

    2024

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.