Pith. sign in

REVIEW 3 major objections 6 minor 28 references

Model Confidence-Guided Multi-Image Fusion of Fundus Images for Diabetic Retinopathy Diagnosis

T0 review · 3 major / 6 minor · reviewed 2026-07-12 · grok-4.5

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

Pith's one-line read Model confidence plus multi-image fusion beats human quality labels for reliable diabetic retinopathy screening from fundus photos.

desk verdict Solid practical result on confidence + multi-image fusion for mobile DR screening, but the headline ~12%/6% gains partly conflate fusion with confidence because the cascade baseline stays single-image. read the letter →

arxiv 2607.03643 v1 pith:GUS4Q23U submitted 2026-07-03 eess.IV cs.CV

Ananya Raghu , Anisha Raghu , Alice S. Tang , Yannis M. Paulus , Tyson N. Kim , Tomiko T. Oskotsky This is my paper · ORCID
classification eess.IVcs.CV
keywords diabeticretinopathyfundusimagingmodelconfidencemulti-imagefusionselectivepredictionmobilescreeningimagequalitypatient-leveldiagnosis
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

This paper argues that for diabetic retinopathy screening from smartphone or clinic fundus images, the model's own prediction confidence is a better filter for unreliable cases than human-annotated image-quality scores. By fusing several images per patient with a lightweight transformer and then accepting only high-confidence patient-level predictions, the method raises balanced accuracy and sensitivity while allowing low-confidence cases to be deferred for a retake. On two Brazilian datasets (mobile and tabletop), the approach substantially outperforms a classic cascade that first discards "poor-quality" images and then diagnoses the rest. The practical payoff is more trustworthy automated screening at capture time in settings where specialists are scarce, using a small model that needs only one forward pass per image.

What carries the argument

Confidence-guided multi-image fusion: a shared lightweight vision backbone embeds each fundus image; a small transformer aggregates the embeddings into one patient-level probability (supporting variable image counts via masking); symmetric confidence margins around 0.5 then decide whether to accept the diagnosis or defer the patient for retake.

What would settle it

On a new multi-image screening cohort from a different population, measure whether a confidence-thresholded fusion model still beats a quality-cascade baseline at matched coverage on balanced accuracy and sensitivity; if the cascade equals or exceeds fusion, or if calibration collapses so that high-confidence bins no longer track true disease rates, the central claim fails.

Watch

Extended reading notes

Core claim

Human quality labels are only weakly linked to actual diagnostic accuracy, whereas confidence-based rejection of patient-level predictions, especially after multi-image fusion, consistently yields higher balanced accuracy and sensitivity at the same coverage rates. At 70 percent patient coverage the fused confidence pipeline reaches roughly 91 percent balanced accuracy on the mobile set and 97 percent on the clinic set, gains of about 12 and 6 points over quality cascade filtering; at 50 percent coverage it reaches 94 percent and 96 percent sensitivity versus 61 percent and 86 percent for the cascade.

Load-bearing premise

The load-bearing premise is that the model's fused probability, even when mildly overconfident, ranks diagnostic reliability well enough that rejecting low-confidence patients improves real screening outcomes.

Editorial extensions

If this is right

  • Screening programs can replace or de-emphasize separate quality classifiers and simply use the diagnostic model's confidence to decide which patients need an immediate retake.
  • Patients kept at high-confidence operating points receive fewer incorrect automated diagnoses, reducing both missed disease and unnecessary referrals.
  • Because the pipeline is lightweight and needs only one inference per image, it can run offline on mobile devices used for field fundus capture.
  • The same confidence-plus-fusion pattern can be applied wherever multiple fundus views of a patient are routinely acquired.

Reading between the lines

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

  • If confidence truly tracks diagnostic utility better than visual quality, many existing "ungradable" rejection rules in retinal AI pipelines may be discarding usable images and lowering overall screening yield.
  • Mild overconfidence may be tolerable for high-sensitivity screening thresholds, but the same system would need recalibration or different margins before high-confidence negatives could safely discharge patients without human review.
  • The masking-friendly transformer suggests a practical path for clinics that capture uneven numbers of images per eye without forcing a fixed two-image binocular architecture.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes a confidence-guided multi-image fusion pipeline for patient-level diabetic retinopathy (DR) screening on fundus photographs. Using a lightweight RETFoundGreen backbone, it (i) fuses multiple images per patient with a small transformer (or mean/max pooling baselines), then (ii) applies a symmetric confidence margin around 0.5 on the fused probability to defer low-confidence cases. On mBRSET (mobile, n=1,234 patients after filtering) and BRSET (clinic, n=7,599), the authors report that at 70% coverage their method reaches ~91% and ~97% balanced accuracy—gains of ~12 and ~6 points over a single-image quality-cascade baseline—and higher sensitivity at matched coverage. They further argue that human image-quality labels are only weakly associated with diagnostic accuracy (Table 6, saliency), so model confidence is a better filter than explicit quality assessment, and that the pipeline is suitable for offline mobile screening.

Significance. If the comparative claims hold under cleaner ablations, the work is practically relevant for LMIC mobile DR screening: it shows that selective deferral plus multi-view aggregation can improve the accuracy–coverage tradeoff without Bayesian multi-pass uncertainty, using a ~25M-parameter offline model and public datasets with released code (SureSight). The independent finding that human quality labels poorly track diagnostic utility (Table 6; occlusion maps) is useful and challenges common cascade designs. Strengths include patient-level splits, mean/max/transformer fusion baselines, coverage-matched operating curves, calibration analysis, and saliency. The main significance risk is attribution: headline gains currently mix multi-image fusion with confidence filtering against a single-image quality cascade, so the incremental value of confidence versus multi-view aggregation is not fully isolated.

major comments (3)
  1. The headline BA/sensitivity gains (Abstract; Results text for Figs. 3–4: ~12%/6% BA at 70% coverage; 94%/96% vs 61%/86% sensitivity at 50% coverage) compare a single-image quality cascade (Methods, Fig. 1a) to confidence-based multi-image fusion (Fig. 2). Tables 3–4 already show multi-image fusion alone lifts BA (mBRSET 0.77→0.84; BRSET 0.90→0.92) without confidence thresholds. Figs. 3–4 therefore conflate multi-view aggregation with confidence-based deferral. Please add matched multi-image quality-cascade (or quality-filtered fusion) and report confidence-only vs fusion-only deltas at fixed coverage so the claim that “confidence-based filtering consistently outperforms image quality-based cascaded pipelines” is cleanly supported.
  2. Fig. 6 shows systematic overconfidence (mean predicted probability exceeds observed DR rate in most bins). Operating points that produce the Abstract numbers are obtained by sweeping the symmetric margin T on the test set (Proposed Framework; Evaluation). For translational claims about “prompting retakes” and reduced incorrect diagnoses, either (a) fix T on validation only and report test metrics once, or (b) apply a simple calibration step (e.g., temperature scaling) and re-evaluate coverage–BA/sensitivity curves. Directional calibration supports ranking, but uncorrected overconfidence weakens the reliability interpretation of high-confidence negatives.
  3. Methods restrict evaluation to patients with fixed image counts (exactly two per eye for mBRSET; one per eye for BRSET; Tables S1–S2), then assign patient labels by max DR across eyes. Real screening has variable numbers of captures and missing eyes. Please quantify selection bias (how many patients/images were dropped and whether DR prevalence differs) and report performance when the transformer is evaluated with naturally incomplete sets (beyond the masking used for two-image mBRSET), or state this as a hard limitation on the coverage claims.
minor comments (6)
  1. Fig. 5 caption: “mage quality” → “Image quality”.
  2. Abstract Methods lists three comparators including “confidence-based prediction,” but Results lead with cascade vs multi-image fusion; make the single-image confidence arm as explicit in the Abstract numbers as in Figs. 3–4.
  3. Equation (1): notation for class weights is hard to parse in the text rendering (w0/w1 definitions); rewrite with standard weighted CE notation.
  4. Tables 3–4 highlight “highest mean” but several metrics are within reported SD of each other; consider marking statistical comparisons or confidence intervals rather than bold alone.
  5. Related work on uncertainty-aware DR (refs 26–27) is cited; a short quantitative note on wall-clock cost vs MC-dropout/ensemble on the same backbone would strengthen the “single inference pass” deployment claim.
  6. BRSET DR prevalence is only 6.4% (Table 2); emphasize that high specificity/BA can be driven by the negative class and report prevalence-sensitive metrics at the chosen operating points more prominently.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: empirical ML evaluation on held-out public data with post-hoc operating curves; results do not reduce to inputs by construction.

full rationale

This is a standard empirical computer-vision paper comparing single-image cascade, confidence filtering, and multi-image transformer fusion pipelines for binary DR detection. All models are trained with weighted cross-entropy on patient-level 70/15/15 splits of the public BRSET and mBRSET datasets; metrics (BA, sensitivity, coverage) are measured on held-out test patients. Confidence thresholds are symmetric margins around 0.5 that are swept after inference solely to generate accuracy-coverage trade-off curves (Figs. 3-4); they are not fitted parameters that force the reported gains. Multi-image aggregation (mean/max/transformer) and selective classification via model probability are conventional techniques whose performance is independently measured, not definitional. There are no self-definitional equations, no uniqueness theorems imported from the authors, no ansatz smuggled via self-citation, and no renaming of known results. The same model supplies both the prediction and its confidence—an inherent feature of selective prediction, not circularity under the listed patterns. The skeptic concern about unmatched cascade vs. fusion baselines is a methodological comparison issue, not a circular derivation. The paper is therefore self-contained against external benchmarks; score 0 with empty steps is the correct outcome.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard deep-learning practice plus two domain choices: that fused model probability is a usable reliability signal, and that the Brazilian multi-image cohorts with fixed image counts adequately represent mobile screening. No new physical entities are invented; free parameters are ordinary training and threshold choices.

free parameters (3)
  • confidence margin T (symmetric around 0.5)
    Swept to produce coverage-accuracy curves; operating points (70% coverage, 50% coverage, 90% sensitivity) are selected post-hoc for reporting.
  • class weights w0, w1 in weighted cross-entropy
    Set from inverse class frequencies to handle imbalance; directly affect training dynamics.
  • transformer fusion depth (two encoder layers) and masking scheme
    Architectural choices that determine how multi-image embeddings are aggregated; not derived from first principles.
assumptions (4)
  • domain assumption Model softmax/probability confidence is a useful ranking signal for selective deferral even when the model is mildly overconfident.
    Invoked throughout the confidence-filtering pipeline and justified only by the calibration plot showing monotonic but overconfident behavior.
  • domain assumption Patient-level DR label can be defined as the maximum ICDR grade across available eye images.
    Stated in Methods; standard but assumes asymmetric disease is correctly captured by max pooling of labels.
  • domain assumption RETFoundGreen pretrained weights provide sufficiently generalizable retinal features for both quality and DR heads.
    All encoders are initialized from this backbone; performance claims inherit its transfer properties.
  • ad hoc to paper Fixed image counts per patient (exactly two per eye for mBRSET, one per eye for BRSET) after exclusion do not introduce selection bias relative to real screening.
    Preprocessing tables S1–S2 discard incomplete patients; the multi-image claim is evaluated only on this restricted cohort.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Model Confidence-Guided Multi-Image Fusion of Fundus Images for Diabetic Retinopathy Diagnosis." pith.science (2026). https://pith.science/paper/GUS4Q23U

@misc{pith2026260703643,
  author       = {Pith},
  title        = {Pith review of: Model Confidence-Guided Multi-Image Fusion of Fundus Images for Diabetic Retinopathy Diagnosis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GUS4Q23U}},
  note         = {Machine review of arXiv:2607.03643}
}
read the original abstract

Purpose: Early screening for eye diseases is critical in low- and middle-income countries where access to care is limited. We investigate whether a confidence-guided, multi-image diabetic retinopathy diagnosis framework can integrate image filtering with confidence-aware predictions for reliable screening at capture. Methods: We develop a multi-image fusion method that aggregates retinal views to improve confidence and balanced accuracy. Our method uses confidence to identify unreliable predictions, prompting retakes when needed. We compare: (1) a cascaded image-quality and disease diagnosis pipeline using a single image per patient, (2) confidence-based prediction, and (3) our confidence-based multi-image fusion pipeline. All methods are evaluated using a RETFoundGreen backbone on the mBRSET (n = 1,234) and BRSET (n = 7,599) datasets. Results: At 70% coverage, our method achieves 91% balanced accuracy on mBRSET and 97% on BRSET, improvements of ~12% and ~6%, respectively, over cascade filtering. The image-quality cascade reaches sensitivities of 61% on mBRSET and 86% on BRSET, whereas our framework reaches 94% and 96%, respectively, at 50% coverage. Conclusions: Human-annotated quality labels are weakly associated with diagnostic performance, and confidence-based filtering consistently outperforms image quality-based cascaded pipelines. Translational Relevance: Using confidence-based multi-image fusion, patients receive more reliable predictions, reducing incorrect diagnoses during screening. The lightweight backbone and single inference pass per image make the framework compatible with low-latency mobile screening systems in resource-limited settings.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 16 canonical work pages

  1. [1]

    Diabetic Retinopathy: A Growing Concern in an Aging Population

    Paulus YM, Gariano RF. Diabetic Retinopathy: A Growing Concern in an Aging Population. Geriatrics. 2009 Feb; 64(2): 16-20. PMID: 19256582

    2009

  2. [2]

    Does the Current Global Health Agenda Lack Vision?

    Forrest SL, Mercado CL, Engmann CM, et al. Does the Current Global Health Agenda Lack Vision?. Glob Health Sci Pract . 2023;11(1):e2200091. Published 2023 Feb 28. doi:10.9745/GHSP-D-22-00091

    work page doi:10.9745/ghsp-d-22-00091 2023

  3. [3]

    Global Prevalence of Diabetic Retinopathy and Projection of Burden through 2045: Systematic Review and Meta-analysis

    Teo Z.L., Tham Y.C., Yu M., Chee M.L., Rim T.H., Cheung N., Bikbov M.M., Wang Y.X., Tang Y., Lu Y., et al. Global Prevalence of Diabetic Retinopathy and Projection of Burden through 2045: Systematic Review and Meta-analysis. Ophthalmology. 2021;128:1580–1591. doi: 10.1016/j.ophtha.2021.04.027

    work page doi:10.1016/j.ophtha.2021.04.027

  4. [4]

    Advances in Structural and Functional Retinal Imaging and Biomarkers for Early Detection of Diabetic Retinopathy

    Zhang Z, Deng C, Paulus YM. Advances in Structural and Functional Retinal Imaging and Biomarkers for Early Detection of Diabetic Retinopathy. Biomedicines 2024 Jun 25, 12(7), 1405. https://doi.org/10.3390/biomedicines12071405 PMID: 39061979; PMCID: PMC11274328. 25

    doi:10.3390/biomedicines12071405 2024

  5. [5]

    Clinical Applications and Future Directions of Smartphone Fundus Imaging

    Ahn SJ, Kim YH. Clinical Applications and Future Directions of Smartphone Fundus Imaging. Diagnostics (Basel) . 2024 Jun 30;14(13):1395. doi:10.3390/diagnostics14131395

    work page doi:10.3390/diagnostics14131395 2024

  6. [6]

    A smartphone-based tool for rapid, portable, and automated wide-field retinal imaging

    Kim TN, Myers F, Reber C, Loury PJ, Loumou P, Webster D, Echanique C, Li P, Davila JR, Maamari RN, Switz NA, Keenan J, Woodward MA, Paulus YM, Margolis T, Fletcher DA. A smartphone-based tool for rapid, portable, and automated wide-field retinal imaging. Transl Vis Sci Technol. 2018 Oct 1; 7(5):21

    2018

  7. [7]

    Smartphone-based, rapid, wide-field fundus photography for diagnosis of pediatric retinal diseases

    Patel TP, Kim TN, Yu G, Dedania VS, Lieu P, Qian CX, Besirli CG, Demirci H, Margolis T, Fletcher DA, Paulus YM. Smartphone-based, rapid, wide-field fundus photography for diagnosis of pediatric retinal diseases. Trans Vis Sci Technol. 2019 May 30; 8(3):29

    2019

  8. [8]

    Smartphone-based fundus photography for screening of plus-disease retinopathy of prematurity

    Patel TP, Aaberg MT, Paulus YM, Lieu P, Dedania VS, Qian CX, Besirli CG, Margolis T, Fletcher DA, Kim TN. Smartphone-based fundus photography for screening of plus-disease retinopathy of prematurity. Graefes Arch Clin Exp Ophthalmol. 2019 Nov; 257(11):2579-2585. doi: 10.1007/s00417-019-04470-4. PMID: 31501929; PMCID: PMC6824990

    work page doi:10.1007/s00417-019-04470-4 2019

Show all 28 references
  1. [9]

    Diabetic retinopathy screening using a portable retinal camera in Vanuatu

    Caceres J, Zhang Y, Boe L, Zhou Y, Besirli C, Paulus YM, Rosenthal J. Diabetic retinopathy screening using a portable retinal camera in Vanuatu. Clinical Ophthalmology. 2023 Oct 4:17, 2919-2927. https://doi.org/10.2147/OPTH.S410425 PMID: 37814638; PMCID: PMC10560479

    2023 doi

  2. [10]

    Comparison of automated and expert human grading of diabetic retinopathy using smartphone-based retinal photography

    Kim TN, Aaberg MT, Li P, Davila JR, Bhaskaranand M, Bhat S, Ramachandra C, Solanki K, Myers F, Reber C, Jalalizadeh R, Margolis TP, Fletcher D, Paulus YM. Comparison of automated and expert human grading of diabetic retinopathy using smartphone-based retinal photography. Eye (...

    2021 doi

  3. [11]

    Automated Quality Assessment of Fundus Images via Analysis of Illumination, Naturalness and Structure,

    F. Shao, Y. Yang, Q. Jiang, G. Jiang and Y. -S. Ho, "Automated Quality Assessment of Fundus Images via Analysis of Illumination, Naturalness and Structure," in IEEE Access, vol. 6, pp. 806-817, 2018, doi: 10.1109/ACCESS.2017.2776126

    2018 doi

  4. [12]

    Retinal image quality assessment using deep learning

    Zago GT, Andreão RV, Dorizzi B, Teatini Salles EO. Retinal image quality assessment using deep learning. Computers in Biology and Medicine . 2018;103:64-70. doi:https://doi.org/10.1016/j.compbiomed.2018.10.004

    2018 doi

  5. [13]

    Quality and content analysis of fundus images using deep learning

    Chalakkal RJ, Abdulla WH, Thulaseedharan SS. Quality and content analysis of fundus images using deep learning. Computers in Biology and Medicine . 2019;108:317-331. doi:https://doi.org/10.1016/j.compbiomed.2019.03.019

    2019 doi

  6. [14]

    DMS-Net:Dual-modal multi-scale siamese network for binocular fundus image classification. (n.d.). https://arxiv.org/html/2504.18046v2

    arXiv

  7. [15]

    mBRSET, a mobile Brazilian retinal dataset [dataset]

    Nakayama LF, Santos F, Barbosa I, Pereira R, Lima R, Oliveira C, et al. mBRSET, a mobile Brazilian retinal dataset [dataset]. PhysioNet. 2024. Available from: https://physionet.org/content/mbrset/1.0/

    2024

  8. [16]

    BRSET: A Brazilian Multilabel Ophthalmological Dataset of Retina Fundus Photos

    Nakayama LF, Pereira RM, Barbosa JHA, Ramos PL, Avila MP, Oliveira MC, et al. BRSET: A Brazilian Multilabel Ophthalmological Dataset of Retina Fundus Photos. PLOS Digit Health . 2024 Jul 11;3(7):e0000454. doi:10.1371/journal.pdig.0000454

    2024 doi

  9. [17]

    P., Ferris, F

    Wilkinson, C. P., Ferris, F. L., Klein, R. E., Lee, P. P., Agardh, C. D., Davis, M., Dills, D., Kampik, A., Pararajasegaram, R., & Verdaguer, J. T. (2003). Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology, 1...

    2003 doi

  10. [18]

    Azad, R., Sinha, S., & Nishant, P. (2021). Asymmetric diabetic retinopathy. Indian journal of ophthalmology, 69(11), 3026–3034. https://doi.org/10.4103/ijo.IJO_1525_21 27

    2021 doi

  11. [19]

    Hubbard LD, Sun W, Cleary PA, et al. Comparison of Digital and Film Grading of Diabetic Retinopathy Severity in the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Study. Arch Ophthalmol. 2011;129(6):718–726. doi:10.1001/archop...

    2011 doi

  12. [20]

    Diabetes Control and Complications Trial Research Group

    Progression of retinopathy with intensive versus conventional treatment in the Diabetes Control and Complications Trial. Diabetes Control and Complications Trial Research Group. (1995). Ophthalmology , 102 (4), 647–661. https://doi.org/10.1016/s0161-6420(95)30973-6

    1995 doi

  13. [21]

    A foundation model for generalizable disease detection from retinal images

    Zhou Y, Chia MA, Wagner SK, et al. A foundation model for generalizable disease detection from retinal images. Nature . 2023;622:156-163. doi:10.1038/s41586-023-06555-X

    2023 doi

  14. [22]

    Training a high-performance retinal foundation model with half-the-data and 400 times less compute

    Engelmann, J., Bernabeu, M.O. Training a high-performance retinal foundation model with half-the-data and 400 times less compute. Nat Commun 16, 6862 (2025). https://doi.org/10.1038/s41467-025-62123-z

    2025 doi

  15. [23]

    Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., & Houlsby, N. (2021). An image is worth 16×16 words: Transformers for image recognition at scale . In International Confer...

    2021

  16. [24]

    P., & Ba, J

    Kingma, D. P., & Ba, J. (2015). Adam: A method for stochastic optimization. In Y. Bengio & Y. LeCun (Eds.), Proceedings of the 3rd International Conference on Learning Representations (ICLR 2015) . International Conference on Learning Representations

    2015

  17. [25]

    DeepFundus: A flow-cytometry-like image quality classifier for boosting the whole life cycle of medical artificial intelligence

    Liu L, Wu X, Lin D, et al. DeepFundus: A flow-cytometry-like image quality classifier for boosting the whole life cycle of medical artificial intelligence. Cell Rep Med . 2023;4(2):100912. doi:10.1016/j.xcrm.2022.100912 28

    2023 doi

  18. [26]

    Uncertainty-Aware Deep Learning Methods for Robust Diabetic Retinopathy Classification,

    J. Jaskari et al., "Uncertainty-Aware Deep Learning Methods for Robust Diabetic Retinopathy Classification," in IEEE Access, vol. 10, pp. 76669-76681, 2022, doi: 10.1109/ACCESS.2022.3192024

    2022 doi

  19. [27]

    Wundram, A.M., Baumgartner, C.F. (2026). Is Uncertainty Quantification a Viable Alternative to Learned Deferral?. In: Sudre, C.H., et al. Uncertainty for Safe Utilization of Machine Learning in Medical Imaging. UNSURE 2025. Lecture Notes in Computer Science, vol 16166. Springe...

    2026 doi

  20. [28]

    Automated Diabetic Retinopathy Detection Based on Binocular Siamese-Like Convolutional Neural Network,

    Dai, L., Wu, L., Li, H. et al. A deep learning system for detecting diabetic retinopathy across the disease spectrum. Nat Commun 12 , 3242 (2021). https://doi.org/10.1038/s41467-021-23458-5 [29] X. Zeng, H. Chen, Y. Luo and W. Ye, "Automated Diabetic Retinopathy Detection Base...

    2021 doi

Pith tools

Reviewed July 12, 2026 · model on record in the stance chip above.