Pith. sign in

REVIEW 4 major objections 7 minor 1 cited by

HOG-CNN: Integrating Histogram of Oriented Gradients with Convolutional Neural Networks for Retinal Image Classification

T0 review · 4 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A hybrid HOG-CNN model fuses handcrafted gradient histograms with frozen pretrained CNN features, and the paper claims this raises retinal disease classification accuracy and AUC over pure deep-learning baselines across three public…

desk verdict Useful appendix, but the headline numbers are a per-dataset best-backbone selection, so the claim that a single HOG-CNN model performs consistently well does not hold as written. read the letter →

arxiv 2507.22274 v1 pith:5ACAMZFG submitted 2025-07-29 cs.CV cs.LG

Faisal Ahmed This is my paper
classification cs.CVcs.LG
keywords RetinaldiseasediagnosisHistogramofOrientedGradientsConvolutionalneuralnetworksFeaturefusionFundusimagesDiabeticretinopathyGlaucomaAge-relatedmaculardegeneration
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

The paper proposes a hybrid classifier, HOG-CNN, that fuses handcrafted Histogram of Oriented Gradients features with features from a frozen pretrained CNN to classify retinal fundus images. It claims this fusion captures both local texture and edge structure and high-level semantic content, improving discrimination for diabetic retinopathy, age-related macular degeneration, and glaucoma over pure deep learning baselines. On APTOS 2019 the model reaches 98.5% accuracy and 99.2 AUC for binary DR and 94.2 AUC for five-class DR; on IC-AMD it reaches 92.8% accuracy and 94.5 AUC; on ORIGA it reaches 83.9% accuracy and 87.2 AUC. The paper argues that because the HOG pathway needs no data augmentation, segmentation, or domain-specific pretraining, the method is a lightweight, interpretable option for resource-constrained screening.

What carries the argument

The central mechanism is a two-branch architecture. A grayscale version of each fundus image is converted into a 26,244-dimensional HOG descriptor with 9 orientation bins, 8x8 cells, and 2x2 normalized blocks, then compressed by a three-layer MLP of sizes 800, 256, and 128; in parallel, the RGB image passes through a frozen pretrained CNN with its classification head removed, followed by a 64-filter 3x3 convolution, 2x2 max pooling, flattening, and a 64-unit dense layer. The two embeddings are concatenated into a 192-dimensional vector and fed to two further dense layers with dropout 0.2 before a sigmoid or softmax output. This design lets the model combine local gradient histograms with deep semantic features.

What would settle it

Run the same four tasks with a single pre-registered backbone and without inspecting test labels; if the HOG+CNN variant does not beat the CNN-only variant in accuracy or AUC on held-out folds, the claimed complementarity is refuted.

Watch

Extended reading notes

Core claim

The central claim is that adding a handcrafted gradient-histogram branch to a deep CNN improves retinal disease classification beyond either feature type alone. In the paper's experiments the fused model beats the CNN-only version on nearly every backbone and dataset combination reported in the appendix, and the headline configuration outperforms published baselines: 98.5% accuracy and 99.2 AUC for binary diabetic retinopathy, 83.2% accuracy with 94.2 AUC for five-class grading, 92.8% accuracy and 94.5 AUC for AMD, and 83.9% accuracy and 87.2 AUC for glaucoma. The author treats this as evidence that local texture information captured by HOG is complementary to the semantic features learned by CNNs in ophthalmic imaging.

Load-bearing premise

The headline results use a different pretrained backbone for each dataset, and the paper never states that those backbone choices were made before seeing test performance, so the claimed consistent advantage could be an artifact of picking the best of several configurations rather than a property of the HOG+CNN fusion.

Editorial extensions

If this is right

  • If the reported results hold under fair comparison, HOG-CNN shows that handcrafted descriptors still add value on top of modern CNN features in medical imaging.
  • The method reaches near-state-of-the-art accuracy without data augmentation, segmentation, or fine-tuning of the backbone, which shortens training and lowers compute requirements.
  • Because the HOG branch is computed in grayscale and the CNN branch is frozen, the model can be trained on CPU and deployed in low-resource clinics.
  • The consistent gains of HOG+CNN over CNN-only in the appendix support the general claim that gradient and texture cues are complementary to deep features across three diseases.
  • The four-task results establish a single fusion recipe that transfers across DR, AMD, and glaucoma without task-specific engineering.

Reading between the lines

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

  • The appendix tests ten backbones with and without HOG, and the headline numbers are the best per dataset; the one-model framing is strongest if each backbone choice was fixed before test performance was inspected.
  • A testable extension is to check whether the HOG gain persists on external or out-of-distribution fundus datasets where illumination and resolution vary; if it does, the method's value is robustness rather than only benchmark accuracy.
  • The reported precision and recall asymmetry on IC-AMD, 94.8% precision versus 72.0% recall, suggests the model is conservative in flagging AMD, so future work could tune the decision threshold to meet clinical recall requirements.
  • The same dual-branch fusion could be combined with other handcrafted descriptors, such as color histograms or topological features, which the paper leaves to future work and which might improve recall on the minority AMD and glaucoma classes.
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 / 7 minor

Summary. The manuscript proposes HOG-CNN, a hybrid architecture that concatenates Histogram of Oriented Gradients (HOG) descriptors with features from a frozen pretrained CNN backbone, and evaluates it on three retinal fundus benchmarks: binary and five-class diabetic retinopathy on APTOS 2019, age-related macular degeneration on IChallenge-AMD, and glaucoma on ORIGA. The main text reports state-of-the-art or near-state-of-the-art accuracy, precision, recall, and AUC numbers for each task, and an appendix tabulates results for ten different CNN backbones with and without HOG fusion. The paper argues that the fusion of handcrafted gradient features with deep features yields consistently high performance while remaining lightweight and interpretable.

Significance. If the reported results were produced by a single, well-specified model under controlled comparison, the paper would provide a useful data point on combining handcrafted HOG features with transfer-learned CNN features for retinal screening, supported by public code and a broad set of per-backbone experiments. The appendix's systematic reporting of ten backbones is a strength. However, the central claim of a single HOG-CNN architecture achieving consistent state-of-the-art performance is not currently supported, because the headline numbers are drawn from different backbones across datasets and the comparisons to prior work are made under incompatible evaluation protocols. The empirical contribution is therefore real but overstated, and the paper in its present form does not support the generalization claims in the abstract and conclusion.

major comments (4)
  1. [Section 5 and Appendix Tables .6-.9] The headline result for each dataset is not produced by a single HOG-CNN configuration. The binary APTOS result in Table 3 corresponds to HOG+EfficientNetB3 in Table .6; the five-class APTOS and IC-AMD results correspond to HOG+EfficientNetB2 in Tables .7 and .8; and the ORIGA result corresponds to HOG+EfficientNetB0 in Table .9. The manuscript does not specify a validation-based rule for selecting these backbones, so the reported numbers are consistent with selecting the best of the ten configurations after observing test performance. This is not a cosmetic issue: in Table .7, plain VGG19 achieves a five-class DR AUC of 95.65, above the reported HOG-CNN AUC of 94.2, and in Table .8, HOG+EfficientNetB3 has an IC-AMD AUC of 93.61 while HOG+EfficientNetB2 has 94.49. The text's claim that a single 'HOG-CNN' model performs consistently at the state of the art is therefore not established by the present experiments.
  2. [Tables 2-5] The comparisons against prior work use different train/test splits or cross-validation schemes across methods: HOG-CNN uses 80:20 for APTOS and ORIGA while many baselines use 85:15 or 10-fold CV; Table 4 lists no split for MobileNetV2-SVM; Table 5 mixes 70:30, 10-fold, and pretrained setups; and the CycleGAN row in Table 2 uses a different class balance (933:267). No confidence intervals or repeated-run variability are reported. Under these conditions, differences of a fraction of a percentage point, which the text interprets as 'surpassing' (e.g., 98.5 vs 98.4 accuracy in Table 3), cannot be distinguished from split-induced variation. The superiority claims in Section 5 require evaluation on identical splits with confidence intervals or significance tests.
  3. [Section 6, first paragraph] The sentence 'the proposed model achieved an accuracy, precision, and recall of 92.8%, along with an AUC of 94.5%' is contradicted by Table 2, which lists precision 94.8 and recall 72.0 for HOG-CNN on the IC-AMD dataset. This factual error also undermines the subsequent argument that HOG-CNN offers 'a more balanced performance across all metrics' than S2C2L, since the reported recall is ten points below S2C2L's 82.4%.
  4. [Section 5, DR Detection Results] In the five-class DR setting, the paper reports precision 86.0, recall 78.6, accuracy 83.2, and AUC 94.2 as the HOG-CNN result. Appendix Table .7 shows that no single backbone simultaneously attains these values: HOG+EfficientNetB0 reaches 83.49 accuracy, while HOG+EfficientNetB2 reaches 94.15 AUC, and the two configurations differ in precision and recall. Reporting a composite of two different configurations as one model's result reinforces the selection-bias concern raised above and should be clarified or corrected.
minor comments (7)
  1. [Highlights and Section 4.1] The dataset is referred to inconsistently as 'IC-Dataset', 'IChallenge-AMD', and 'IC-AMD'; one name should be used throughout.
  2. [Section 3.2] The text describing the HOG pathway reads 'the first dense layer contains 800 ReLU-activated units, the second dense layer contains 256 ReLU-activated units, and the second contains 128 ReLU-activated units'; the last occurrence should be 'the third'.
  3. [Algorithm 1 reference] The phrase 'summarized in here 1' should be 'summarized in Algorithm 1'.
  4. [Table .8] In the InceptionResNetV2 row, the precision value '2152' appears to be a typo for 21.52.
  5. [Table 5] The CNN-SVM row lists a class balance of 660:453 for ORIGA, although the dataset is described as containing 650 images; this discrepancy needs clarification.
  6. [General reporting] AUC values are reported with percent signs (e.g., 99.2%), which is nonstandard for a unitless quantity; either report AUC as a decimal or state the convention explicitly.
  7. [Section 2] The related-work section contains an extended discussion of topological data analysis that is not connected to the proposed method or experiments, making the section harder to follow.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: empirical benchmark evaluation; headline-backbone variation is a selection/reporting concern, not a derivation that reduces to its inputs.

full rationale

This is an empirical benchmark paper rather than a theoretical derivation, so the circularity burden is low. HOG features are computed by the standard Dalal-Triggs descriptor, CNN embeddings come from frozen pretrained backbones, and the fusion is a fixed concatenation followed by dense layers. The reported accuracies and AUCs are measured on external public benchmarks (APTOS, ORIGA, IC-AMD) under stated train/test protocols; no equation defines the model in terms of the target result, and no parameter is fitted to a headline metric and then renamed as a prediction. The author self-citations in the Related Work (Refs. [12,13,14]) are background TDA references and are not load-bearing for the HOG-CNN claims. The appendix shows that the headline results use different pretrained backbones across datasets and that some plain backbones outperform the reported HOG-CNN figures on individual metrics; this is a possible model-selection and reporting concern, but it is not circularity, because the reported numbers are not forced to equal their inputs by construction. No uniqueness theorem, ansatz-smuggling citation, or renamed known result carries the argument. The derivation chain, such as it is, is self-contained against external benchmarks, so the appropriate circularity finding is no significant circularity.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The reported results depend on several hand-chosen and potentially post hoc choices, especially the per-dataset backbone selection. No new entities are introduced; the assumptions are standard domain assumptions for transfer learning and HOG feature use.

free parameters (6)
  • Backbone selection per dataset = EfficientNetB3 (binary APTOS), EfficientNetB2 (5-class APTOS and IC-AMD), EfficientNetB0 (ORIGA)
    Appendix Tables 6-9 show different backbones are best for each dataset, yet the main tables report a single HOG-CNN. This indicates post hoc selection from 10 candidates, which is a hidden choice that affects all results.
  • HOG cell size = 8x8 pixels
    Hand-chosen standard HOG parameter; no ablation or optimization reported.
  • HOG block size = 2x2 cells
    Hand-chosen standard HOG parameter; no ablation.
  • HOG orientation bins = 9
    Hand-chosen standard HOG parameter; no ablation.
  • Dense layer sizes and dropout = 800, 256, 128; 256, 128; dropout 0.2
    Network structure chosen by hand; no justification or hyperparameter search described.
  • Batch size and training epochs = 32; 50
    Chosen by hand; Adam optimizer defaults left unchanged.
assumptions (4)
  • domain assumption Pretrained ImageNet CNN features transfer to retinal fundus images.
    The model freezes the CNN backbone without fine-tuning (Section 3.2).
  • domain assumption HOG descriptors carry discriminative information for retinal disease classification.
    The paper assumes gradient orientation features help separate DR, AMD, and glaucoma (Section 3.1).
  • domain assumption No pixel normalization is needed for the pretrained CNN.
    Section 3.2 states 'No normalization is applied to pixel intensities', which is atypical for ImageNet-pretrained models and is not justified.
  • domain assumption Comparing models across different train/test splits is fair.
    The paper compares HOG-CNN (80:20 split) against baselines using 85:15 or 10-fold splits and treats the comparison as valid (Tables 2-5).

how reviews work

0 comments
Cite this review

Pith. "Pith review of HOG-CNN: Integrating Histogram of Oriented Gradients with Convolutional Neural Networks for Retinal Image Classification." pith.science (2026). https://pith.science/paper/5ACAMZFG

@misc{pith2026250722274,
  author       = {Pith},
  title        = {Pith review of: HOG-CNN: Integrating Histogram of Oriented Gradients with Convolutional Neural Networks for Retinal Image Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5ACAMZFG}},
  note         = {Machine review of arXiv:2507.22274}
}
read the original abstract

The analysis of fundus images is critical for the early detection and diagnosis of retinal diseases such as Diabetic Retinopathy (DR), Glaucoma, and Age-related Macular Degeneration (AMD). Traditional diagnostic workflows, however, often depend on manual interpretation and are both time- and resource-intensive. To address these limitations, we propose an automated and interpretable clinical decision support framework based on a hybrid feature extraction model called HOG-CNN. Our key contribution lies in the integration of handcrafted Histogram of Oriented Gradients (HOG) features with deep convolutional neural network (CNN) representations. This fusion enables our model to capture both local texture patterns and high-level semantic features from retinal fundus images. We evaluated our model on three public benchmark datasets: APTOS 2019 (for binary and multiclass DR classification), ORIGA (for Glaucoma detection), and IC-AMD (for AMD diagnosis); HOG-CNN demonstrates consistently high performance. It achieves 98.5\% accuracy and 99.2 AUC for binary DR classification, and 94.2 AUC for five-class DR classification. On the IC-AMD dataset, it attains 92.8\% accuracy, 94.8\% precision, and 94.5 AUC, outperforming several state-of-the-art models. For Glaucoma detection on ORIGA, our model achieves 83.9\% accuracy and 87.2 AUC, showing competitive performance despite dataset limitations. We show, through comprehensive appendix studies, the complementary strength of combining HOG and CNN features. The model's lightweight and interpretable design makes it particularly suitable for deployment in resource-constrained clinical environments. These results position HOG-CNN as a robust and scalable tool for automated retinal disease screening.

Figures

Figures reproduced from arXiv: 2507.22274 by the authors.

Figure 1
Figure 1. Illustration of the preprocessing pipeline applied to a fundus image. From left to right: (a) Original color image, (b) grayscale conversion [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed model HOG-CNN architecture: The model accepts two inputs: (a) a color fundus image and (f) a grayscale fundus image. The top pathway processes (a) through (b) a pre-trained CNN, (c) a 64-filter convolutional layer, (d) a 2 × 2 max-pooling layer, and (e) a flattening layer. In the bottom pathway, (g) the HOG visualization is extracted from (f), and (h) its feature vector is passed through (i)… view at source ↗
Figure 3
Figure 3. Training and Testing Accuracy and Loss over 50 Epochs on the APTOS Dataset (Binary Classification). [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Accuracy comparison of the proposed model with existing deep learning methods: (a) and (b) show bar plots for binary and 5-class [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

Discussion (0). Sign in 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. RepViT-CXR: A Channel Replication Strategy for Vision Transformers in Chest X-ray Tuberculosis and Pneumonia Classification

    cs.CV 2025-09 reject novelty 2.0 of 10

    Applying a standard channel-replication trick to a pretrained Vision Transformer yields very high reported accuracy for TB and pneumonia classification on three chest X-ray datasets.

Reference graph

Works this paper leans on

61 extracted references · 59 canonical work pages · cited by 1 Pith paper

  1. [1]

    W. H. Organization, Blindness and vision im- pairment, 2023. URL: https://www.who. int/news-room/fact-sheets/detail/ blindness-and-visual-impairment , accessed: 2025-05-02

    work page 2023

  2. [2]

    P. S. Silva, D. Zhang, C. M. P. Jacoba, W. Fickweiler, D. Lewis, J. Leitmeyer, K. Curran, R. P. Salongcay, D. Doan, M. Ashraf, et al., Automated machine learning for predicting diabetic retinopathy progression from ultra-widefield retinal images, JAMA ophthalmology 142 (2024) 171–178

    work page 2024

  3. [3]

    E. R. Dow, H. K. Jeong, E. A. Katz, C. A. Toth, D. Wang, T. Lee, D. Kuo, M. J. Allingham, M. Hadziahmetovic, P. S. Mettu, et al., A deep-learning algorithm to predict short-term progression to geographic atrophy on spectral-domain optical coherence to- mography, JAMA ophthalmology 141 (2023) 1052–1061

    work page 2023

  4. [4]

    T. Li, W. Bo, C. Hu, H. Kang, H. Liu, K. Wang, H. Fu, Appli- cations of deep learning in fundus images: A review, Medical Image Analysis 69 (2021) 101971

    work page 2021

  5. [5]

    D. S. Ting, et al., Deep learning in ophthalmology: the technical and clinical considerations, Progress in retinal and eye research 72 (2019) 100759

    work page 2019

  6. [6]

    M. H. Sarhan, et al., Machine learning techniques for oph- thalmic data processing: a review, IEEE Journal of Biomedical and Health Informatics 24 (2020) 3338–3350

    work page 2020

  7. [7]

    D. S. W. Ting, et al., Artificial intelligence and deep learning in ophthalmology, British Journal of Ophthalmology 103 (2019) 167–175

    work page 2019

  8. [8]

    Dalal, B

    N. Dalal, B. Triggs, Histograms of oriented gradients for hu- man detection, in: 2005 IEEE computer society conference on computer vision and pattern recognition (CVPR’05), volume 1, Ieee, 2005, pp. 886–893

    work page 2005

Show all 61 references
  1. [9]

    Singh, S

    A. Singh, S. Sengupta, V . Lakshminarayanan, Explainable deep learning models in medical image analysis, Journal of imaging 6 (2020) 52

    2020

  2. [10]

    Fourcade, R

    A. Fourcade, R. H. Khonsari, Deep learning in medical image analysis: A third eye for doctors, Journal of stomatology, oral and maxillofacial surgery 120 (2019) 279–288

    2019

  3. [11]

    Pratt, F

    H. Pratt, F. Coenen, D. M. Broadbent, S. P. Harding, Y . Zheng, Convolutional neural networks for diabetic retinopathy, Proce- dia computer science 90 (2016) 200–205

    2016

  4. [12]

    Ahmed, M

    F. Ahmed, M. A. N. Bhuiyan, B. Coskunuzer, Topo-cnn: Reti- nal image analysis with topological deep learning, Journal of Imaging Informatics in Medicine (2025) 1–17

    2025

  5. [13]

    Ahmed, B

    F. Ahmed, B. Coskunuzer, Tofi-ml: Retinal image screening with topological machine learning, in: Annual Conference on Medical Image Understanding and Analysis, Springer, 2023, pp. 281–297

    2023

  6. [14]

    Ahmed, Topological Machine Learning in Medical Image Analysis, Ph.D

    F. Ahmed, Topological Machine Learning in Medical Image Analysis, Ph.D. thesis, The University of Texas at Dallas, 2023

    2023

  7. [15]

    Qaiser, Y .-W

    T. Qaiser, Y .-W. Tsang, D. Taniyama, N. Sakamoto, K. Nakane, D. Epstein, N. Rajpoot, Fast and accurate tumor segmentation of histology images using persistent homology and deep convo- lutional features, Medical image analysis 55 (2019) 1–14

    2019

  8. [16]

    Lawson, A

    P. Lawson, A. B. Sholl, J. Q. Brown, B. T. Fasy, C. Wenk, Per- sistent homology for the quantitative evaluation of architectural features in prostate cancer histology, Scientific reports 9 (2019) 1139

    2019

  9. [17]

    Yadav, F

    A. Yadav, F. Ahmed, O. Daescu, R. Gedik, B. Coskunuzer, Histopathological cancer detection with topological signatures, in: 2023 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2023, pp. 1610–1619. doi: 10.1109/ BIBM58861.2023.10385822

    2023

  10. [18]

    Berry, Y .-C

    E. Berry, Y .-C. Chen, J. Cisewski-Kehe, B. T. Fasy, Functional summaries of persistence diagrams, Journal of Applied and Computational Topology 4 (2020) 211–262

    2020

  11. [19]

    Crawford, et

    L. Crawford, et. al., Predicting clinical outcomes in glioblas- toma: an application of topological and functional data analysis, J.Amer.Stat.Assoc. 115 (2020) 1139–1150

    2020

  12. [20]

    Yadav, F

    A. Yadav, F. Ahmed, O. Daescu, R. Gedik, B. Coskunuzer, Histopathological cancer detection with topological signatures, in: 2023 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), IEEE, 2023, pp. 1610–1619

    2023

  13. [21]

    Ahmed, B

    F. Ahmed, B. Nuwagira, F. Torlak, B. Coskunuzer, Topo-CXR: Chest X-ray TB and Pneumonia Screening with Topological Machine Learning, in: Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision, 2023, pp. 2326–2336

    2023

  14. [22]

    Kanari, et al., A topological representation of branching neu- ronal morphologies, Neuroinformatics 16 (2018) 3–13

    L. Kanari, et al., A topological representation of branching neu- ronal morphologies, Neuroinformatics 16 (2018) 3–13

    2018

  15. [23]

    Bendich, J

    P. Bendich, J. S. Marron, E. Miller, A. Pieloch, S. Skwerer, Per- sistent homology analysis of brain artery trees, The annals of applied statistics 10 (2016) 198

    2016

  16. [24]

    Rieck, et al., Uncovering the topology of time-varying fmri data using cubical persistence, NeurIPS 33 (2020) 6900–6912

    B. Rieck, et al., Uncovering the topology of time-varying fmri data using cubical persistence, NeurIPS 33 (2020) 6900–6912

    2020

  17. [25]

    B. J. Stolz, et al., Topological data analysis of task-based fmri data from experiments on schizophrenia, Journal of Physics: Complexity 2 (2021) 035006

    2021

  18. [26]

    P. G. C ´amara, A. J. Levine, R. Rabadan, Inference of ancestral recombination graphs through topological data analysis, PLoS computational biology 12 (2016) e1005071

    2016

  19. [27]

    J. I. Orlando, et al., An ensemble deep learning based approach for red lesion detection in fundus images, Computer methods and programs in biomedicine 153 (2018) 115–127

    2018

  20. [28]

    A. You, J. K. Kim, I. H. Ryu, T. K. Yoo, Application of gen- erative adversarial networks (gan) for ophthalmology image do- mains: a survey, Eye and Vision 9 (2022) 1–19

    2022

  21. [29]

    Srivastava, et.al., Artificial intelligence and machine learning in ophthalmology: A review, Indian Journal of Ophthalmology 71 (2023) 11–17

    O. Srivastava, et.al., Artificial intelligence and machine learning in ophthalmology: A review, Indian Journal of Ophthalmology 71 (2023) 11–17

    2023

  22. [30]

    N. D. Koseoglu, T. Liu, Predictive deep learning applications in ophthalmology, Rev. Ophthalmol 17 (2023)

    2023

  23. [31]

    Ashtari-Majlan, M

    M. Ashtari-Majlan, M. M. Dehshibi, D. Masip, Glaucoma di- agnosis in the era of deep learning: A survey, Expert Systems with Applications 256 (2024) 124888

    2024

  24. [32]

    D. M. Nguyen, H. M. T. Alam, T. Nguyen, D. Srivastav, H.- J. Profitlich, N. Le, D. Sonntag, Deep learning for ophthal- mology: The state-of-the-art and future trends, arXiv preprint arXiv:2501.04073 (2025)

    2025 arXiv

  25. [33]

    Fang, et al., ADAM challenge: Detecting age-related mac- ular degeneration from fundus images, IEEE Transactions on Medical Imaging (2022)

    H. Fang, et al., ADAM challenge: Detecting age-related mac- ular degeneration from fundus images, IEEE Transactions on Medical Imaging (2022)

    2022

  26. [34]

    Fu, et al., ADAM: Automatic detection challenge on AMD,

    H. Fu, et al., ADAM: Automatic detection challenge on AMD,

  27. [35]

    Z. Zhang, et al., Origa-light: An online retinal fundus image database for glaucoma analysis and research, in: 2010 Annual international conference of the IEEE engineering in medicine and biology, IEEE, 2010, pp. 3065–3068

    2010

  28. [36]

    https://www.kaggle.com/c/ aptos2019-blindness-detection

    APTOS, Asia Pacific Tele-Ophthalmology So- ciety (APTOS) 2019 Blindness Detection Dataset, 2019. https://www.kaggle.com/c/ aptos2019-blindness-detection

    2019

  29. [37]

    Goutam, M

    B. Goutam, M. F. Hashmi, Z. W. Geem, N. D. Bokde, A com- prehensive review of deep learning strategies in retinal disease diagnosis using fundus images, IEEE Access (2022)

    2022

  30. [38]

    M. R. Islam, et al., Applying supervised contrastive learning for the detection of DR and its severity levels from fundus images, Computers in Biology and Medicine 146 (2022) 105602

    2022

  31. [39]

    J. D. Bodapati, et. al., Composite deep neural network with gated-attention mechanism for DR severity classification, J. 14 Amb. Int. Hum. Compt. 12 (2021) 9825–9839

    2021

  32. [40]

    Macsik, J

    P. Macsik, J. Pavlovicova, J. Goga, S. Kajan, Local binary cnn for diabetic retinopathy classification on fundus images, Acta Polytech. Hung. 19 (2022) 27–45

    2022

  33. [41]

    A. R. Singh, G. Singh, N. Saluja, L. Garg, Mobilenetv2-svm hybrid model classification of diabetic retinopathy effected hu- man eye, in: 2024 Global Conference on Communications and Information Technologies (GCCIT), IEEE, 2024, pp. 1–5

    2024

  34. [42]

    L. K. Singh, M. Khanna, H. Garg, R. Singh, M. Iqbal, A three- stage novel framework for efficient and automatic glaucoma classification from retinal fundus images, Multimedia Tools and Applications 83 (2024) 85421–85481

    2024

  35. [43]

    Latif, S

    J. Latif, S. Tu, C. Xiao, S. Ur Rehman, A. Imran, Y . Latif, Odgnet: a deep learning model for automated optic disc local- ization and glaucoma classification using fundus images, SN Applied Sciences 4 (2022) 1–11

    2022

  36. [44]

    Liao, et al., Clinical interpretable deep learning model for glaucoma diagnosis, IEEE journal of biomedical and health in- formatics 24 (2019) 1405–1412

    W. Liao, et al., Clinical interpretable deep learning model for glaucoma diagnosis, IEEE journal of biomedical and health in- formatics 24 (2019) 1405–1412

    2019

  37. [45]

    Elangovan, M

    P. Elangovan, M. K. Nath, Glaucoma assessment from color fundus images using convolutional neural network, Int. J. of Imaging Systems and Technology 31 (2021) 955–971

    2021

  38. [46]

    X. Zhao, F. Guo, Y . Mai, J. Tang, X. Duan, B. Zou, L. Jiang, Glaucoma screening pipeline based on clinical measurements and hidden features, IET Image Processing 13 (2019) 2213– 2223

    2019

  39. [47]

    Ajitha, J

    S. Ajitha, J. D. Akkara, M. Judy, Identification of glaucoma from fundus images using deep learning techniques, Indian Journal of Ophthalmology 69 (2021) 2702

    2021

  40. [48]

    Z. Wu, Y . Xiong, S. X. Yu, D. Lin, Unsupervised feature learning via non-parametric instance discrimination, in: CVPR, 2018, pp. 3733–3742

    2018

  41. [49]

    M. Ye, X. Zhang, P. C. Yuen, S.-F. Chang, Unsupervised em- bedding learning via invariant and spreading instance feature, in: CVPR, 2019, pp. 6210–6219

    2019

  42. [50]

    Z. Feng, C. Xu, D. Tao, Self-supervised representation learning by rotation feature decoupling, in: CVPR, 2019, pp. 10364– 10374

    2019

  43. [51]

    T. Chen, S. Kornblith, M. Norouzi, G. Hinton, A simple frame- work for contrastive learning of visual representations, in: ICML, PMLR, 2020, pp. 1597–1607

    2020

  44. [52]

    X. Li, X. Hu, X. Qi, L. Yu, W. Zhao, P.-A. Heng, L. Xing, Rotation-oriented collaborative self-supervised learning for reti- nal disease diagnosis, IEEE Transactions on Medical Imaging 40 (2021) 2284–2294

    2021

  45. [53]

    Zhang, Z

    Z. Zhang, Z. Ji, Q. Chen, S. Yuan, W. Fan, Joint optimization of cyclegan and cnn classifier for detection and localization of retinal pathologies on color fundus photographs, IEEE Journal of Biomedical and Health Informatics 26 (2021) 115–126

    2021

  46. [54]

    Chakraborty, ., A

    R. Chakraborty, ., A. Pramanik, Dcnn-based prediction model for detection of AMD from color fundus images, Medical & Bio. Eng. & Comp. 60 (2022) 1431–1448

    2022

  47. [55]

    Q. Bi, H. Zheng, X. Sun, J. Yi, W. Zhang, Y . Huang, Y . Li, Y . Zheng, Self-supervised cross-level consistency learning for fundus image classification, in: ICASSP 2024-2024 IEEE Inter- national Conference on Acoustics, Speech and Signal Process- ing (ICASSP), IEEE, 2024, pp....

    2024

  48. [56]

    S. S. Chaturvedi, K. Gupta, V . Ninawe, P. S. Prasad, Automated diabetic retinopathy grading using deep convolutional neural network, arXiv preprint arXiv:2004.06334 (2020)

    2020 arXiv

  49. [57]

    J. D. Bodapati, et. al., Blended multi-modal deep convnet fea- tures for diabetic retinopathy severity prediction, Electronics 9 (2020) 914

    2020

  50. [58]

    Kumar, S

    G. Kumar, S. Chatterjee, C. Chattopadhyay, Dristi: a hybrid deep neural network for diabetic retinopathy diagnosis, Signal, Image and Video Processing 15 (2021) 1679–1686

    2021

  51. [59]

    F. Long, H. Xiong, J. Sang, A classification method for diabetic retinopathy based on self-supervised learning, in: International Conference on Intelligent Computing, Springer, 2024, pp. 347– 357

    2024

  52. [60]

    I. A. Taj, et al., An ensemble framework based on deep CNNs for glaucoma classification., Mathematical Biosciences and En- gineering 18 (2021) 5321–5347. 15

    2021

  53. [2020]

    doi:10.21227/dt4f-rt59

    doi

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.