Pith. sign in

REVIEW 2 major objections 5 minor 31 references

GlaKG: A Biomarker-Centric Fundus Knowledge Graph for Explainable Glaucoma Diagnosis and Risk Assessment

T0 review · 2 major / 5 minor · reviewed 2026-07-11 · grok-4.5

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

Pith's one-line read GlaKG gives every glaucoma diagnosis an explicit chain of biomarker evidence and clinical rules.

desk verdict Solid, honest methods paper: a concrete fundus KG schema plus post-hoc fusion that actually ships rule-level audit trails, with metrics correctly framed as upper bounds under clean biomarkers. read the letter →

arxiv 2607.04673 v1 pith:QU2O3MSN submitted 2026-07-06 cs.CV cs.LG

Cheng Huang , Jia Zhang , Yi Jiang , Yang Liu , Karanjit Kooner , Yadi Liu , Tsengdar Lee , Yang Xie
show 2 more authors
Wenqi Shi Guanghua Xiao
This is my paper
classification cs.CVcs.LG
keywords GlaucomaDiagnosisKnowledgeGraphExplainableAIRiskStratificationNeuralNetworkFundusImagingClinicalRulesBiomarkers
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

Most automated glaucoma systems score fundus images with deep networks but leave the decision opaque. This paper builds GlaKG, a knowledge graph that encodes structural biomarkers, eleven established clinical rules, and image features so that every prediction comes with a readable reasoning chain. A ResNet50 image score and a normalized rule-chain score are fused only at inference, with all fitting kept on the training split. On a public AI-annotated fundus set the method reaches near-ceiling binary F1 and strong four-class risk accuracy; the authors openly call these figures an upper bound because the annotations are highly label-correlated. The claim that matters for practice is not raw accuracy alone but that clinicians can see which rules fired, at what strength, and that low chain scores flag borderline cases instead of silent failures.

What carries the argument

GlaKG’s reasoning-chain score: a weighted sum of activated clinical rules (CDR thresholds, ISNT violation, rim thinning, bayoneting, etc.) that is min-max normalized and fused with the image probability as p_final = (1−α)p_img + α s_KG, with α chosen by training-split cross-validation only.

What would settle it

Apply the same fusion pipeline on a fundus cohort whose biomarkers are measured independently of the diagnosis labels (or extracted automatically from raw images) and test whether chain scores still rank cases and lift risk-stratification accuracy over image-only baselines.

Watch

Extended reading notes

Core claim

A biomarker-centric fundus knowledge graph with six entity types, eight relation types, and eleven clinically validated rules can produce, for every sample, an explicit reasoning chain from biomarker evidence to activated rules; post-processing fusion of ResNet50 probability with the normalized chain score then yields F1 0.9953 for binary glaucoma classification and 0.930 accuracy with 0.922 weighted F1 for four-class risk stratification on an AI-annotated public dataset, framed as an upper bound attainable with clean structured biomarkers rather than leakage-free image-only performance.

Load-bearing premise

The method needs high-quality structured biomarker fields—like the AI annotations where pathological signs cleanly separate glaucoma from normal—so the fixed clinical rules can fire and produce a meaningful chain score.

Editorial extensions

If this is right

  • Every automated glaucoma call can carry a clinician-readable audit trail of which biomarkers and rules drove it.
  • Borderline cases become visible as low chain scores rather than silent misclassifications.
  • Image encoders and clinical rule sets can be updated independently without end-to-end retraining.
  • KG-derived and biomarker features contribute near-equally (51.1% vs 48.9%), showing structured knowledge is complementary to visual cues.
  • The same schema supports four-class risk stratification with higher accuracy and weighted F1 than image-only baselines.

Reading between the lines

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

  • If automatic biomarker extractors from raw fundus images reach similar fidelity, the audit trail could transfer to clinics without pre-made JSON annotations.
  • The modular split between fixed rules and image features suggests the same pattern could encode rule sets for diabetic retinopathy or macular degeneration.
  • Low chain-score flagging could be wired into referral workflows so only uncertain cases escalate to specialists.
  • Cross-device and cross-dataset tests would show whether the fixed clinical thresholds remain stable when annotation quality varies.
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

2 major / 5 minor

Summary. The paper introduces GlaKG, a biomarker-centric heterogeneous knowledge graph for explainable glaucoma diagnosis and risk stratification from fundus images. It encodes six entity types, eight relation types, and 11 clinically grounded rules so that each prediction is accompanied by an explicit reasoning chain from biomarker attributes to activated rules (Eqs. 6–9). Image evidence (ResNet50 embeddings + Gradient Boosting) is combined with a min-max-normalized chain score via post-processing convex fusion (Eq. 12), with all fitting confined to the training split. On a public AI-annotated dataset the method reports F1 = 0.9953 for binary classification and 0.930 accuracy / 0.922 weighted F1 for four-class risk stratification, framed as upper bounds under clean structured biomarkers rather than leakage-free image-only performance. Ablations on fusion weight α, GNN baselines, feature-importance analysis, and a single false-negative case (CDR = 0.700, low chain score) support the interpretability claim.

Significance. If the framework holds, it supplies a concrete, clinician-auditable alternative to saliency-map XAI for glaucoma: every decision exposes which rules fired, at what strength, and how they aggregate, including low-score flagging of borderline cases. The pipeline hygiene (original split, train-only PCA/standardization/α selection, s_KG never used to train the image classifier) and open disclosure of annotation–label correlation are strengths that make the upper-bound framing credible. The modular design (image branch separable from rule anchors) is a practical contribution for incremental clinical integration. The work is therefore significant as an interpretability architecture even when headline metrics are understood as contingent on high-quality structured biomarkers.

major comments (2)
  1. Sec. IV-A/B and Tables III, VI, VII: the reported F1/accuracy rest on AI-generated biomarker fields that separate classes almost perfectly (100% rim thinning and ISNT violation in glaucoma, 0% pathological signs in normals; KG-only AUC = 1.000 at α = 1). The paper correctly frames these numbers as upper bounds and lists structured-JSON dependence as a limitation (Sec. VII), yet the abstract and contribution bullets still lead with the headline metrics. For the central claim to transfer, the manuscript should either (i) add an experiment with imperfect/noisy biomarkers (or automatic extraction) that degrades the chain score in a controlled way, or (ii) demote the numerical claims further so that the primary result is the leakage-invariant reasoning-chain audit trail rather than the upper-bound F1.
  2. Eqs. 9–12 and Sec. V-C: the chain score CS is a fixed weighted sum of rule activations on the same attributes that define the labels on this dataset. Fusion with an independent image probability is a reasonable post-processing design, but the near-equal MDI contribution of KG vs. biomarker features (51.1% vs. 48.9%, Table V) is partly circular when both are derived from the same JSON fields. A clearer separation—e.g., reporting image-only vs. pure-rule performance on a second dataset without label-correlated annotations, or an ablation that removes the most label-correlated rules—would strengthen the claim that the KG supplies non-redundant clinical structure rather than re-encoding the annotation schema.
minor comments (5)
  1. Fig. 1 caption states α* = 0.5 while the text correctly notes that α* is selected by 5-fold CV; make the caption consistent with the selection procedure.
  2. Table VII reports average rules/chain scores over all 689 diagnoses (including normals), which are lower than the TP subset means in Table VI; a short clarifying sentence already exists but could be moved earlier to avoid reader confusion.
  3. The four-class AUC for the fused model (Table II) does not exceed the strongest image-only baseline; the text notes this but could briefly discuss why accuracy/weighted-F1 improve while AUC does not (minority-class operating-point effects).
  4. Notation: ClinicalRule nodes are excluded from |Tv| = 6 yet appear in the total node count; a single sentence in Sec. III-A1 already explains this, but the schema figure caption could restate it for self-containment.
  5. Minor typography: “▷ ◁k” in Eq. (6) and occasional missing spaces around α in the fusion equation; standard copy-editing will fix these.

Circularity Check

1 steps flagged · score 5.0 of 10

KG chain score (and its near-perfect ranking) is a deterministic function of the dataset’s label-correlated biomarker annotations via fixed clinical rules; paper openly frames the resulting metrics as an upper bound rather than independent discovery.

  1. fitted input called prediction [Sec. III-B (Eqs. 6–9), Sec. IV-B1, Table III (α=1.0 row), Sec. V-C]
    "The reasoning chain score aggregates weighted rule activations: CS(v_diag)=∑ s_k · 1[r_k,v_b] … At α=1.0, the KG score ranks cases near-perfectly (AUC=1.000) … This confirms that the KG signal’s separating power derives from label-correlated biomarkers (Sec. IV-B1) … the fused F1=0.9953 is therefore an upper bound achievable with clean structured biomarkers, not a claim of leakage-free image-only performance."

    CS is a fixed, deterministic weighted sum of rule indicators evaluated on the supplied biomarker attributes. Those attributes are nearly perfectly class-separating by the dataset’s own construction (100 % prevalence of key signs in glaucoma, 0 % in normals). Hence the perfect ranking of s_KG (and the large lift when it is fused) is statistically forced by the input annotations rather than being an independent prediction; the paper correctly labels the numbers an upper bound, but the reduction of the KG-only “prediction” to its inputs remains by construction.

full rationale

The paper’s core technical claim is an auditable post-processing fusion of ResNet50 image scores with a KG reasoning-chain score CS defined by 11 fixed, guideline-derived rules (Eqs. 6–9) applied to structured biomarker fields. All fitting (PCA, standardization, Gradient Boosting, α selection) is confined to the training split and s_KG never enters training, so there is no training-time leakage or self-citation load-bearing for the rules themselves (rules cite external clinical literature). However, on this particular AI-annotated dataset the biomarker fields separate classes almost perfectly (100 % rim thinning / ISNT violation in glaucoma, 0 % pathological signs in normals). Consequently the normalized s_KG ranks with AUC = 1.000 by construction of the input annotations rather than by any independent visual or statistical inference; the fused headline numbers are therefore forced once the clean structured biomarkers are supplied. The authors explicitly disclose this correlation and re-frame the metrics as an upper bound attainable with clean biomarkers, and the interpretability contribution (explicit rule activations and low-score flagging of borderline cases) remains leakage-invariant. The circularity is therefore partial and confined to the performance numbers under the given annotations; it does not collapse the framework claim itself. No self-citation uniqueness theorems, ansatz smuggling, or renaming of known results are load-bearing.

Assumptions & free parameters 4 free parameters · 4 assumptions · 2 invented entities

The central interpretability claim rests on clinical rules and thresholds taken as given from ophthalmic literature, a fixed strength scale {1,2,3}, a hand-chosen entity/relation schema, and a fusion weight selected on training F1. Predictive numbers further rest on AI-generated biomarker JSON that the paper itself treats as label-correlated. No new physical entity is postulated; the invented pieces are the graph schema and the chain-score aggregator.

free parameters (4)
  • fusion weight alpha* = 0.5
    Chosen by 5-fold CV over {0.0,0.3,0.5,0.7,1.0} to maximize F1 on the training split; reported optimum 0.5 drives the headline fused metrics.
  • rule strength weights s_k = {1,2,3} per rule
    Each of the 11 rules is assigned weak/moderate/strong weight in {1,2,3} from cited clinical literature; these discrete weights directly define CS and the normalized s_KG.
  • PCA embedding dimension = 64
    ResNet50 2048-d features are projected to 64-d for classifiers and GNN node features; dimension is a modeling choice fit on train embeddings.
  • decision threshold on p_final = 0.5
    Binary label is 1 iff p_final >= 0.5 (Eq. 13); fixed operating point interacts with alpha and explains the alpha=1.0 precision/recall tradeoff.
assumptions (4)
  • domain assumption Eleven clinical rules with fixed comparison operators and thresholds (e.g., CDR >= 0.7, ISNT violation, rim thinning) correctly encode actionable glaucoma diagnostic criteria.
    Sec. III-B formalizes rules as 4-tuples from ophthalmic guidelines [24–27]; thresholds are not refit to labels but are still assumed sufficient for the reasoning chain.
  • domain assumption AI-generated structured biomarker JSON fields are faithful enough proxies for clinical biomarkers to evaluate reasoning chains.
    Sec. IV-A/B uses the public AI-annotated dataset; authors note strong label correlation, so this axiom is load-bearing for any transfer claim.
  • ad hoc to paper Post-processing convex fusion of an image probability and a min-max normalized chain score is a valid way to combine visual and knowledge evidence without end-to-end joint training.
    Eqs. 10–13 define the fusion; alpha is tuned on train F1 rather than derived from a calibrated probabilistic model.
  • standard math Standard GCN/GAT message-passing and ResNet50 ImageNet features are adequate visual/graph encoders for this task.
    Sec. III-D and IV use off-the-shelf architectures with conventional training settings.
invented entities (2)
  • GlaKG heterogeneous schema (6 entity types, 8 relation types, ClinicalRule anchors)
    purpose: Unify fundus images, structural biomarkers, pathologies, diagnoses, risk levels, and rule supports into one typed graph for reasoning chains.
    Defined in Sec. III-A and Fig. 2; schema is paper-specific engineering, not an independently measured biological object.
  • Reasoning chain score CS and normalized s_KG
    purpose: Aggregate weighted rule activations into a single [0,1] knowledge score for fusion and uncertainty flagging.
    Eqs. 9–11 invent the aggregator (max attainable 18 under mutually compatible rules); falsifiable only insofar as low CS flags borderline cases on this dataset.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GlaKG: A Biomarker-Centric Fundus Knowledge Graph for Explainable Glaucoma Diagnosis and Risk Assessment." pith.science (2026). https://pith.science/paper/QU2O3MSN

@misc{pith2026260704673,
  author       = {Pith},
  title        = {Pith review of: GlaKG: A Biomarker-Centric Fundus Knowledge Graph for Explainable Glaucoma Diagnosis and Risk Assessment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QU2O3MSN}},
  note         = {Machine review of arXiv:2607.04673}
}
read the original abstract

Glaucoma is a leading cause of irreversible blindness worldwide, yet most automated diagnosis systems rely on opaque deep-learning models that offer little clinical interpretability. We present GlaKG, a biomarker-centric fundus knowledge graph that integrates structural biomarkers, clinically grounded rules, and image features to produce traceable reasoning for glaucoma diagnosis and risk stratification. GlaKG encodes six entity types (Fundus Image, Optic Disc, Neural Rim, Pathology, Diagnosis, Risk Level), eight relation types, and 11 clinically validated rules into a unified graph, so that every prediction is accompanied by an explicit reasoning chain linking biomarker evidence to activated clinical rules. To keep knowledge-based reasoning strictly separate from label information, we adopt a post-processing fusion framework that combines ResNet50 image embeddings with a normalized KG reasoning-chain score via a tunable weight alpha, with all fitting confined to the training split. On a publicly available, AI-annotated fundus dataset, GlaKG reaches F1 = 0.9953 for binary glaucoma classification and 0.930 accuracy with 0.922 weighted F1 for four-class risk stratification; we report openly that the dataset's biomarker annotations are highly label-correlated, and therefore frame these figures as an upper bound attainable with clean structured biomarkers rather than as leakage-free image-only performance. Feature-importance analysis shows KG-derived and biomarker features contributing near-equally (51.1% vs. 48.9%), and the reasoning chain flags borderline cases by exposing low chain scores rather than failing silently. GlaKG's central contribution is therefore a clinically auditable reasoning framework that complements raw predictive performance by explicitly exposing the biomarker evidence and rule activations behind each decision.

Figures

Figures reproduced from arXiv: 2607.04673 by the authors.

Figure 2
Figure 2. GlaKG entity-relation schema. Node shapes denote entity types: circles [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 1
Figure 1. End-to-end GlaKG framework. The image branch (top) extracts [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Reasoning chain visualization for a representative high-risk glaucoma [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Feature importance (MDI) for glaucoma classification. Blue bars [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

31 extracted references

  1. [1]

    Global prevalence of glaucoma and projections of glaucoma burden through 2040: A systematic review and meta-analysis,

    Y .-C. Tham, X. Li, T. Y . Wong, H. A. Quigley, T. Aung, and C.-Y . Cheng, “Global prevalence of glaucoma and projections of glaucoma burden through 2040: A systematic review and meta-analysis,”Oph- thalmology, vol. 121, no. 11, pp. 2081–2090, 2014

  2. [2]

    Efficacy of a deep learning system for detecting glaucomatous optic neuropathy based on color fundus photographs,

    Z. Li, Y . He, S. Keel, W. Meng, R. T. Chang, and M. He, “Efficacy of a deep learning system for detecting glaucomatous optic neuropathy based on color fundus photographs,”Ophthalmology, vol. 125, no. 8, pp. 1199–1206, 2018

    2018

  3. [3]

    Explainable AI for glaucoma detection and classifica- tion: A comprehensive review,

    S. Akteret al., “Explainable AI for glaucoma detection and classifica- tion: A comprehensive review,”Artificial Intelligence Review, 2025

    2025

  4. [4]

    Learning a health knowledge graph from electronic medical records,

    M. Rotmensch, Y . Halpern, A. Tlimat, S. Horng, and D. Sontag, “Learning a health knowledge graph from electronic medical records,” Scientific Reports, vol. 7, no. 1, p. 5994, 2017

    2017

  5. [5]

    Graph convolutional transformer: Learning the graphical structure of electronic health records,

    E. Choi, Z. Xu, Y . Li, M. Dusenberry, G. Flores, Y . Xue, and A. Dai, “Graph convolutional transformer: Learning the graphical structure of electronic health records,” inProceedings of the AAAI Conference on Artificial Intelligence, 2020, pp. 606–613

    2020

  6. [6]

    REFUGE challenge: A unified framework for evaluating automated methods for glaucoma assessment from fundus photographs,

    J. I. Orlandoet al., “REFUGE challenge: A unified framework for evaluating automated methods for glaucoma assessment from fundus photographs,”Medical Image Analysis, vol. 59, p. 101570, 2020

    2020

  7. [7]

    X-gan: A generative ai-powered unsupervised model for main vessel segmentation of glaucoma screening,

    C. Huang, W. Xie, T. Lee, K. Kooner, J.-K. Wang, N. Zhang, and J. Zhang, “X-gan: A generative ai-powered unsupervised model for main vessel segmentation of glaucoma screening,” inInternational Conference on Neural Information Processing. Springer, 2025, pp. 67–81

    2025

  8. [8]

    AI-driven biomarker discovery & progression modeling for precision diagnosis of glaucoma,

    C. Huang, “AI-driven biomarker discovery & progression modeling for precision diagnosis of glaucoma,” PhD Dissertation, Southern Methodist University, 2026, computer Science and Engineering Theses and Disser- tations, No. 53

    2026

Show all 31 references
  1. [9]

    GlaBoost: Boosting glaucoma detection with multi- modal biomarker integration,

    C. Huanget al., “GlaBoost: Boosting glaucoma detection with multi- modal biomarker integration,” inProceedings of the IEEE Engineering in Medicine and Biology Conference (EMBC), 2026

    2026

  2. [10]

    A systematic review of AI for predicting glaucoma progression: Challenges and recommendations towards clinical imple- mentation,

    Whiteet al., “A systematic review of AI for predicting glaucoma progression: Challenges and recommendations towards clinical imple- mentation,”npj Digital Medicine, vol. 9, p. 140, 2026

    2026

  3. [11]

    Artificial intelligence in the prediction of glaucoma development and progression: A systematic review,

    W. Y . L. Yanget al., “Artificial intelligence in the prediction of glaucoma development and progression: A systematic review,”Survey of Ophthalmology, vol. 71, pp. 235–247, 2025

    2025

  4. [12]

    Glalstm: A concurrent lstm stream framework for glaucoma detection via biomarker relationship mining,

    C. Huang, W. Xie, T. Lee, K. Kooner, N. Zhang, and J. Zhang, “Glalstm: A concurrent lstm stream framework for glaucoma detection via biomarker relationship mining,” in2025 47th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2025

    2025

  5. [13]

    Modeling polypharmacy side effects with graph convolutional networks,

    M. Zitnik, M. Agrawal, and J. Leskovec, “Modeling polypharmacy side effects with graph convolutional networks,”Bioinformatics, vol. 34, no. 13, pp. i457–i466, 2018

    2018

  6. [14]

    Pre-training of graph aug- mented transformers for medication recommendation,

    J. Shang, C. Xiao, T. Ma, H. Li, and J. Sun, “Pre-training of graph aug- mented transformers for medication recommendation,” inProceedings of the International Joint Conference on Artificial Intelligence (IJCAI), 2019, pp. 3953–3959

    2019

  7. [15]

    Human disease ontology 2018 update: Classifica- tion, content and workflow expansion,

    L. M. Schrimlet al., “Human disease ontology 2018 update: Classifica- tion, content and workflow expansion,”Nucleic Acids Research, vol. 47, no. D1, pp. D955–D962, 2019

    2018

  8. [16]

    Protein interface prediction using graph convolutional networks,

    A. Fout, J. Byrd, B. Shariat, and A. Ben-Hur, “Protein interface prediction using graph convolutional networks,” inAdvances in Neural Information Processing Systems (NeurIPS), 2017, pp. 6530–6539

    2017

  9. [17]

    Temporal graph neural network-powered paper recommen- dation on dynamic citation networks,

    J. Shen, M. A. A. Haqqani, B. Hu, C. Huang, X. Xie, T. Lee, and J. Zhang, “Temporal graph neural network-powered paper recommen- dation on dynamic citation networks,”AAAI Workshop on Scientific Document Understanding (SDU@AAAI-2024), 2024

    2024

  10. [18]

    Convolutional graph isomorphism network to detect glaucomatous visual field defects,

    D. R. da Costaet al., “Convolutional graph isomorphism network to detect glaucomatous visual field defects,”Ophthalmology Science, vol. 6, no. 3, p. 101041, 2025

    2025

  11. [19]

    Graph attention networks,

    P. Veli ˇckovi´c, G. Cucurull, A. Casanova, A. Romero, P. Li`o, and Y . Ben- gio, “Graph attention networks,” inProceedings of the International Conference on Learning Representations (ICLR), 2018

    2018

  12. [20]

    Semi-supervised classification with graph convolutional networks,

    T. N. Kipf and M. Welling, “Semi-supervised classification with graph convolutional networks,” inProceedings of the International Conference on Learning Representations (ICLR), 2017

    2017

  13. [21]

    Grad-CAM: Visual explanations from deep networks via gradient-based localization,

    R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad-CAM: Visual explanations from deep networks via gradient-based localization,” inProceedings of the IEEE International Conference on Computer Vision (ICCV), 2017, pp. 618–626

    2017

  14. [22]

    GraphXAI: A survey of graph neural networks (GNNs) for explainable AI (XAI),

    M. Nandan, S. Mitra, and D. De, “GraphXAI: A survey of graph neural networks (GNNs) for explainable AI (XAI),”Neural Computing and Applications, vol. 37, pp. 10 949–11 000, 2025

    2025

  15. [23]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770–778

    2016

  16. [24]

    Optic nerve damage in human glaucoma: II. the site of injury and susceptibility to damage,

    H. A. Quigley, E. M. Addicks, W. R. Green, and A. E. Maumenee, “Optic nerve damage in human glaucoma: II. the site of injury and susceptibility to damage,”Archives of Ophthalmology, vol. 99, no. 4, pp. 635–649, 1981

    1981

  17. [25]

    The optic disc in glaucoma: Pathogenetic correlations,

    G. L. Spaeth, “The optic disc in glaucoma: Pathogenetic correlations,” Transactions of the American Ophthalmological Society, vol. 79, pp. 619–674, 1981

    1981

  18. [26]

    Ophthalmoscopic evaluation of the optic nerve head,

    J. B. Jonas, W. M. Budde, and S. Panda-Jonas, “Ophthalmoscopic evaluation of the optic nerve head,”Survey of Ophthalmology, vol. 43, no. 4, pp. 293–320, 1999

    1999

  19. [27]

    The ISNT rule and differentiation of normal from glaucomatous eyes,

    N. Harizman, C. Oliveira, A. Chiang, C. Tello, J. M. Liebmann, R. Ritch, and K. Mehta, “The ISNT rule and differentiation of normal from glaucomatous eyes,”Archives of Ophthalmology, vol. 124, no. 11, pp. 1579–1583, 2006

    2006

  20. [28]

    Glaucoma diagnosis JSON analysis dataset,

    A. C. Manoj, “Glaucoma diagnosis JSON analysis dataset,” https://huggingface.co/datasets/AswanthCManoj/glaucoma diagnosis json analysis, 2024, accessed: 2025

    2024

  21. [29]

    The regression analysis of binary sequences,

    D. R. Cox, “The regression analysis of binary sequences,”Journal of the Royal Statistical Society: Series B (Methodological), vol. 20, no. 2, pp. 215–232, 1958

    1958

  22. [30]

    Random forests,

    L. Breiman, “Random forests,”Machine Learning, vol. 45, no. 1, pp. 5–32, 2001

    2001

  23. [31]

    Learning repre- sentations by back-propagating errors,

    D. E. Rumelhart, G. E. Hinton, and R. J. Williams, “Learning repre- sentations by back-propagating errors,”Nature, vol. 323, no. 6088, pp. 533–536, 1986

    1986

Pith tools

Reviewed July 11, 2026 · model on record in the stance chip above.