Pith. sign in

REVIEW 3 major objections 5 minor 54 references

SwinECAT: A Transformer-based fundus disease classification model with Shifted Window Attention and Efficient Channel Attention

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read This paper claims that adding an efficient channel attention module after each stage of a Swin Transformer reaches 88.29% accuracy on nine-class fundus disease classification, the highest reported on the EDID dataset.

desk verdict A straightforward Swin+ECA combination on a 9-class fundus benchmark; the 1.7-point gain over Swin is plausible but the paper's single-split, fixed-hyperparameter evaluation leaves the 'significantly outperforms' claim unsupported. read the letter →

arxiv 2507.21922 v1 pith:5NPNZE2W submitted 2025-07-29 cs.CV cs.AI

classification cs.CVcs.AI
keywords fundusdiseaseclassificationSwinTransformerefficientchannelattentionshiftedwindowmedicalimageEDIDdatasetnine-class
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

Paper claim: grafting an Efficient Channel Attention module onto each stage of a Swin Transformer produces the best published nine-class fundus-disease classifier on the EDID benchmark, at 88.29% accuracy, macro F1 of 0.90, and weighted F1 of 0.88. The proposed SwinECAT model is designed to tackle two well-known difficulties in fundus images: lesions that occupy tiny regions and diseases whose visual differences are subtle. Because the ECA block adds only about 0.77M parameters over the 27.53M Swin baseline, the paper argues that channel attention is a cheap and effective complement to shifted-window spatial attention. If the result holds, it means a modest architectural change to a standard transformer can move the needle on a fine-grained medical classification task.

What carries the argument

The model's central object is the Swin Transformer backbone, a hierarchical four-stage vision transformer whose attention is computed inside local windows and shifted between consecutive layers to allow cross-window information exchange, augmented with one Efficient Channel Attention (ECA) block after each stage. ECA compresses each feature map by global average pooling, applies a one-dimensional convolution with an adaptively chosen kernel size along the channel dimension, and uses a sigmoid to produce per-channel weights that rescale the features. The argument is that the shifted-window attention supplies multi-scale spatial modeling from local details to global context, while ECA, with only a tiny parameter cost, steers the network toward the most informative channels, so the two mechanisms jointly improve discrimination in fundus images.

What would settle it

Retrain Swin Transformer and SwinECAT from scratch on the same EDID split across multiple random seeds and compute the distribution of test accuracy; the central claim of an ECA-driven improvement collapses if the accuracy distributions overlap substantially (e.g., the 1.73-point gap is within one standard deviation).

Watch

Extended reading notes

Core claim

The central discovery is that fusing the Swin Transformer backbone with ECA modules, one after each of its four stages, improves classification performance on a nine-class fundus disease task: 88.29% accuracy, macro F1 of 0.9000, and weighted F1 of 0.8830, versus 86.56% accuracy and macro F1 of 0.8849 for the Swin Transformer baseline, with about 0.77M added parameters. The paper reports that this accuracy is higher than six general vision baselines (ViT, ConViT, MaxViT, BEiT, ResNet50, Swin Transformer) and two recent fundus-specific models (MaxViT+ResNet18 and CNN-Transformer), while using fewer parameters than both recent fundus-specific models. The authors attribute the gain to the complementary fusion of spatial attention from shifted windows and channel attention from ECA, which makes the model more sensitive to small lesion regions and fine inter-class differences.

Load-bearing premise

The load-bearing premise is that every compared model is trained to a representative performance level under the shared hyperparameters (learning rate 1e-5, batch size 32, early stopping with patience 3), so the reported gaps reflect architecture differences rather than undertrained baselines.

Editorial extensions

If this is right

  • A nine-class fundus screening system can reach roughly 88% accuracy with about 28 million parameters, which is lighter than the comparable MaxViT+ResNet fusion model.
  • Adding ECA to Swin Transformer yields a 1.7 percentage point accuracy gain with a minimal parameter increase, suggesting channel re-weighting is a high-leverage modification for medical image classification.
  • Pure global-attention transformers such as ViT, BEiT, and ConViT, and heavyweight fusion models such as CNN-Transformer, are less competitive than the hybrid spatial-channel design on this task.
  • The same four-stage SwinECAT recipe is a plausible starting point for other small-lesion medical imaging tasks like diabetic retinopathy grading or optic-disc analysis.

Reading between the lines

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

  • Beyond the paper: a single 80/10/10 split underlies every number in Table 3, so the 1.73-point gain over Swin Transformer should be checked for seed-to-seed variance before treating it as stable.
  • Beyond the paper: placing ECA inside each Swin block rather than after each stage, or combining it with class-balanced sampling for the smallest classes, would isolate whether the gain comes from channel attention or from the extra processing at stage boundaries.
  • Beyond the paper: because EDID was captured with two Topcon cameras in one hospital network, external validation on other fundus datasets would test whether the channel-attention advantage carries across devices and populations.
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 / 5 minor

Summary. The paper proposes SwinECAT, a modification of the Swin Transformer that inserts Efficient Channel Attention (ECA) modules into each of the four hierarchical stages, and evaluates it on the EDID fundus dataset for 9-category disease classification. The authors report 88.29% accuracy, macro F1 0.90, and weighted F1 0.88 on a single 80/10/10 split, claiming that these results significantly outperform the Swin Transformer baseline and other general-purpose and fundus-specific models, and that this is the highest reported accuracy for this task on this dataset.

Significance. If the empirical claims are correct, the paper would provide a simple, parameter-efficient enhancement of a widely used vision transformer backbone for a clinically relevant 9-class medical image classification task, with a modest gain over the Swin Transformer baseline. The use of a public dataset and the inclusion of parameter counts for all compared models are strengths. However, the significance is substantially tempered by the lack of statistical rigor: all results come from a single data split with no repeated runs, no standard deviations, and no significance testing, and the baseline comparisons appear to rest on fixed hyperparameters without evidence that each model was at a representative operating point. As presented, the central claim of 'significant' improvement and state-of-the-art performance is not adequately supported, although the underlying architecture is reasonable and the claimed direction of improvement is plausible.

major comments (3)
  1. [§4.1, §4.2, Table 3] The claim that SwinECAT 'significantly outperform[s]' the baselines is not supported by the experimental protocol. All results are from a single 80/10/10 split with no repeated runs, no standard deviations, and no statistical significance test. The reported gain over the Swin Transformer baseline is 1.73 percentage points (88.29 vs 86.56), which could easily be within run-to-run variance for deep models on a test set of 1,614 images. The paper should report results across multiple seeds or cross-validation folds, and should apply an appropriate significance test (e.g., a paired test across seeds) before using the word 'significantly'.
  2. [§4.1, Table 3] The baseline comparison is confounded by the fixed hyperparameter protocol. The paper states that all models use learning rate 1e-5, batch size 32, and early stopping with patience 3, but it does not state whether each model was tuned for this task or which pretrained weights (if any) were used. The CNN-Transformer baseline achieving only 79.93% accuracy, far below what its components (Inception-ResNet-v2 plus ViT) typically achieve on fundus tasks, strongly suggests undertraining or suboptimal initialization. Without evidence that every baseline was at a representative operating point—for example, by tuning learning rate and early stopping per model or by showing consistent results across a small hyperparameter sweep—the comparative claim that SwinECAT outperforms these baselines is not established.
  3. [Abstract, §4.2.1, Table 3] The state-of-the-art claim ('the highest reported performance for 9-category classification on this public dataset') is not substantiated within the manuscript. The comparison is limited to the authors' own re-implementations of a small set of baselines; there is no systematic comparison with published results on the same EDID dataset, and no discussion of whether those published results used the same data split, preprocessing, or class definitions (e.g., whether pterygium was excluded). To support the state-of-the-art claim, the authors should compare against existing published numbers on this dataset under a consistent evaluation protocol, or at least clearly state that no such comparison was performed.
minor comments (5)
  1. [§3.2.1] The sentence 'On this basis, ECA module is added after the four stages' is ambiguous: it could mean after each stage or only after the final stage. Given Figure 1 and the surrounding text, it likely means 'after each of the four stages,' but the wording should be corrected for reproducibility.
  2. [§4.2.1] In the paragraph discussing Figure 3(g-i), the sentence comparing SwinECAT's validation loss with 'the combined model method of Figure 3(g) and Figure 3(i)' appears to contain a typo: the comparison should be with Figure 3(g) and Figure 3(h), not with itself. Please verify and correct.
  3. [§4.1] The experimental setup should explicitly state whether pretrained weights were used for each model and, if so, which checkpoint (e.g., ImageNet-1k or ImageNet-21k). This detail is essential for reproducibility and for interpreting the baseline results.
  4. [Table 3] The table formatting is difficult to read because the column headers are run together (e.g., 'Parameters(M)ViT' instead of a separate column). Please format the table so that each column is clearly separated.
  5. [References] There is a duplicated reference in the introduction: '[15, 15, 16]' should likely be '[14, 15, 16]' or similar. Also, there are several minor grammatical and punctuation issues throughout (e.g., inconsistent capitalization of 'SwinECAT'/'swinECAT' and 'Transformer' in places like 'Swin T ransformer').

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: SwinECAT is an empirical architecture comparison with no fitted-parameter derivation, self-citation chain, or defined-in-terms-of-itself result.

full rationale

The paper makes no claim that could reduce to its own inputs by construction. SwinECAT is proposed by inserting an ECA module after each stage of the Swin Transformer, and its evaluation is a direct benchmark on the EDID dataset. There is no fitted parameter that is later renamed as a prediction; the reported accuracy and F1 scores are measured outputs, not quantities determined by a fitting procedure. The ablation experiment compares SwinECAT with the Swin Transformer baseline under identical hyperparameters, which is a standard empirical test of a module addition, not a circular step. All cited works (Swin Transformer, ECA-Net, ViT, BEiT, MaxViT, ConViT, ResNet, and the two fundus-disease baselines) are external publications by other authors; there are no load-bearing self-citations, no imported uniqueness theorems, and no ansatz smuggled in through the authors' own prior work. The concern raised by the skeptical reviewer—that baselines may be undertrained because fixed learning rate, batch size, and early-stopping patience were applied uniformly across architectures—is a legitimate threat to external validity and to the strength of the 'significantly outperforms' claim, but it is not a circularity. It concerns whether the comparison is fair, not whether any prediction is equivalent to the model's inputs by definition. Accordingly, the appropriate circularity score is 0.

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

This empirical paper adds no fitted physical constants and no invented entities. The central claim rests on the chosen training hyperparameters (learning rate, batch size, early stopping patience, input size), the validity of the EDID labels and split, and the assumption that the same hyperparameters give every baseline a fair chance.

free parameters (4)
  • learning_rate = 1e-5
    Chosen by hand in Section 4.1 and kept identical for all models, which may disadvantage some architectures.
  • batch_size = 32
    Chosen in Section 4.1 for all models; fixed across the comparison.
  • early_stopping_patience = 3 epochs
    Section 4.1; a short patience may truncate training differently for different models, affecting the fairness of the comparison.
  • input_size = 224x224 center crop
    Section 3.1.2: short side resized to 256, then center cropped to 224. A standard choice, but still a manual selection.
assumptions (3)
  • domain assumption The EDID dataset labels for the 9 categories are correct and the single 80/10/10 split is representative.
    Sections 3.1 and 4.1. Label noise or an unrepresentative split would invalidate all reported numbers.
  • domain assumption All baseline models attain a representative performance level under the shared hyperparameters and unspecified pretrained initialization.
    Section 4.1. This fairness assumption is the foundation of the 'outperforms baselines' claim.
  • standard math The standard implementations of Swin Transformer (reference [23]) and ECA (reference [24]) behave as described in the cited papers.
    The model quality depends on these external implementations, which are not shipped in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SwinECAT: A Transformer-based fundus disease classification model with Shifted Window Attention and Efficient Channel Attention." pith.science (2026). https://pith.science/paper/5NPNZE2W

@misc{pith2026250721922,
  author       = {Pith},
  title        = {Pith review of: SwinECAT: A Transformer-based fundus disease classification model with Shifted Window Attention and Efficient Channel Attention},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5NPNZE2W}},
  note         = {Machine review of arXiv:2507.21922}
}
read the original abstract

In recent years, artificial intelligence has been increasingly applied in the field of medical imaging. Among these applications, fundus image analysis presents special challenges, including small lesion areas in certain fundus diseases and subtle inter-disease differences, which can lead to reduced prediction accuracy and overfitting in the models. To address these challenges, this paper proposes the Transformer-based model SwinECAT, which combines the Shifted Window (Swin) Attention with the Efficient Channel Attention (ECA) Attention. SwinECAT leverages the Swin Attention mechanism in the Swin Transformer backbone to effectively capture local spatial structures and long-range dependencies within fundus images. The lightweight ECA mechanism is incorporated to guide the SwinECAT's attention toward critical feature channels, enabling more discriminative feature representation. In contrast to previous studies that typically classify fundus images into 4 to 6 categories, this work expands fundus disease classification to 9 distinct types, thereby enhancing the granularity of diagnosis. We evaluate our method on the Eye Disease Image Dataset (EDID) containing 16,140 fundus images for 9-category classification. Experimental results demonstrate that SwinECAT achieves 88.29\% accuracy, with weighted F1-score of 0.88 and macro F1-score of 0.90. The classification results of our proposed model SwinECAT significantly outperform the baseline Swin Transformer and multiple compared baseline models. To our knowledge, this represents the highest reported performance for 9-category classification on this public dataset.

Figures

Figures reproduced from arXiv: 2507.21922 by the authors.

Figure 1
Figure 1. The overall structure of the SwinECAT model. It mainly consists of four stages, each of [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. The key components of the SwinECAT model. Figure 2(a) shows the structure of the Swin [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Loss function and accuracy of SwinECAT and other comparison models during training [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 51 canonical work pages

  1. [12]

    Automated detection of nine infantile fundus diseases and conditions in retinal images using a deep learning system

    Yaling Liu et al. “Automated detection of nine infantile fundus diseases and conditions in retinal images using a deep learning system”. In: EPMA Journal 15.1 (2024), pp. 39–51

  2. [1]

    Addressing social determinants of vision health

    Andrew M Williams and Jos´ e-Alain Sahel. “Addressing social determinants of vision health”. In: Ophthalmology and Therapy11.4 (2022), pp. 1371–1382

  3. [2]

    Disparities in vision health and eye care

    Angela R Elam et al. “Disparities in vision health and eye care”. In: Ophthalmology 129.10 (2022), e89–e113

  4. [3]

    Artificial intelligence and deep learning in ophthalmology

    Daniel Shu Wei Ting et al. “Artificial intelligence and deep learning in ophthalmology”. In: British Journal of Ophthalmology103.2 (2019), pp. 167–175

  5. [4]

    Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs

    Varun Gulshan et al. “Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs”. In: jama 316.22 (2016), pp. 2402–2410

  6. [5]

    A Comprehensive Survey on Detection of Ocular and Non-Ocular Diseases Using Color Fundus Images

    Megha Gupta et al. “A Comprehensive Survey on Detection of Ocular and Non-Ocular Diseases Using Color Fundus Images”. In: IEEE Access(2024)

  7. [6]

    Advances in medical image analysis with vision transformers: a comprehensive review

    Reza Azad et al. “Advances in medical image analysis with vision transformers: a comprehensive review”. In: Medical Image Analysis91 (2024), p. 103000

  8. [7]

    Automated identification and grading system of diabetic retinopathy using deep neural networks

    Wei Zhang et al. “Automated identification and grading system of diabetic retinopathy using deep neural networks”. In: Knowledge-Based Systems 175 (2019), pp. 12–25

Show all 54 references
  1. [8]

    Machine Learning Approaches in High Myopia: Systematic Review and Meta- Analysis

    Huiyi Zuo et al. “Machine Learning Approaches in High Myopia: Systematic Review and Meta- Analysis”. In: Journal of Medical Internet Research27 (2025), e57644

  2. [9]

    Transformers in medical imaging: A survey

    Fahad Shamshad et al. “Transformers in medical imaging: A survey”. In: Medical image analysis 88 (2023), p. 102802

  3. [10]

    Comparison of vision transformers and convolutional neural networks in medical image analysis: a systematic review

    Satoshi Takahashi et al. “Comparison of vision transformers and convolutional neural networks in medical image analysis: a systematic review”. In: Journal of Medical Systems 48.1 (2024), p. 84

  4. [11]

    A Review of CNN and Transformer Applications in Image Process- ing

    Zhu Cheng and Niu Xiaoqin. “A Review of CNN and Transformer Applications in Image Process- ing”. In: 2024 9th International Conference on Intelligent Informatics and Biomedical Sciences (ICIIBMS). Vol. 9. IEEE. 2024, pp. 590–596

  5. [13]

    Lesion identification in fundus images via convolutional neural network-vision transformer

    Jian Lian and Tianyu Liu. “Lesion identification in fundus images via convolutional neural network-vision transformer”. In: Biomedical Signal Processing and Control88 (2024), p. 105607

  6. [14]

    Automatic feature learning to grade nuclear cataracts based on deep learning

    Xinting Gao, Stephen Lin, and Tien Yin Wong. “Automatic feature learning to grade nuclear cataracts based on deep learning”. In: IEEE Transactions on Biomedical Engineering 62.11 (2015), pp. 2693–2701

  7. [15]

    Diagnosis of diabetic retinopathy using deep neural networks

    Zhentao Gao et al. “Diagnosis of diabetic retinopathy using deep neural networks”. In: IEEE Access 7 (2018), pp. 3360–3370

  8. [16]

    Deep learning-based fully automated grading system for dry eye disease severity

    Seonghwan Kim et al. “Deep learning-based fully automated grading system for dry eye disease severity”. In: PLoS One 19.3 (2024), e0299776

  9. [17]

    DIA-VXNET: A framework for automated diabetic eye disease detection using transfer learning with feature fusion network

    Md Najib Hasan et al. “DIA-VXNET: A framework for automated diabetic eye disease detection using transfer learning with feature fusion network”. In:Biomedical Signal Processing and Control 100 (2025), p. 106907

  10. [18]

    DeepDiabetic: an identification system of diabetic eye diseases using deep neural networks

    Arwa Albelaihi and Dina M Ibrahim. “DeepDiabetic: an identification system of diabetic eye diseases using deep neural networks”. In: IEEE Access12 (2024), pp. 10769–10789

  11. [19]

    A Novel Multi-Modal Deep Learning Framework for Early Detection of Ocular Diseases

    D Ranjith and M Sakthivanitha. “A Novel Multi-Modal Deep Learning Framework for Early Detection of Ocular Diseases”. In: 2025 International Conference on Intelligent Computing and Control Systems (ICICCS). IEEE. 2025, pp. 1340–1346

  12. [20]

    Challenges for ocular disease identifica- tion in the era of artificial intelligence

    Neha Gour, Muhammad Tanveer, and Pritee Khanna. “Challenges for ocular disease identifica- tion in the era of artificial intelligence”. In: Neural Computing and Applications35.31 (2023), pp. 22887–22909

  13. [21]

    Review of eye dis- eases detection and classification using deep learning techniques

    Ahmed Aizaldeen Abdullah, Ahmed Aldhahab, and Hanaa M Al Abboodi. “Review of eye dis- eases detection and classification using deep learning techniques”. In: BIO Web of Conferences. Vol. 97. EDP Sciences. 2024, p. 00012. 15

  14. [22]

    Classification of eye diseases in fundus images

    Omar Bernab´ e et al. “Classification of eye diseases in fundus images”. In: IEEE Access9 (2021), pp. 101267–101276

  15. [23]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu et al. “Swin transformer: Hierarchical vision transformer using shifted windows”. In: Proceedings of the IEEE/CVF international conference on computer vision. 2021, pp. 10012– 10022

  16. [24]

    ECA-Net: Efficient channel attention for deep convolutional neural net- works

    Qilong Wang et al. “ECA-Net: Efficient channel attention for deep convolutional neural net- works”. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020, pp. 11534–11542

  17. [25]

    The advance of deep learning and attention mechanism

    Xi Chen. “The advance of deep learning and attention mechanism”. In: 2022 International Con- ference on Electronics and Devices, Computational Science (ICEDCS). IEEE. 2022, pp. 318– 321

  18. [26]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy et al. “An image is worth 16x16 words: Transformers for image recognition at scale”. In: arXiv preprint arXiv:2010.11929(2020)

  19. [27]

    Convit: Improving vision transformers with soft convolutional inductive biases

    St´ ephane d’Ascoli et al. “Convit: Improving vision transformers with soft convolutional inductive biases”. In: International conference on machine learning. PMLR. 2021, pp. 2286–2296

  20. [28]

    Beit: Bert pre-training of image transformers

    Hangbo Bao et al. “Beit: Bert pre-training of image transformers”. In: arXiv preprint arXiv:2106.08254 (2021)

  21. [29]

    Maxvit: Multi-axis vision transformer

    Zhengzhong Tu et al. “Maxvit: Multi-axis vision transformer”. In: European conference on com- puter vision. Springer. 2022, pp. 459–479

  22. [30]

    Deep residual learning for image recognition

    Kaiming He et al. “Deep residual learning for image recognition”. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2016, pp. 770–778

  23. [31]

    A dataset of color fundus images for the detection and classification of eye diseases

    Shayla Sharmin et al. “A dataset of color fundus images for the detection and classification of eye diseases”. In: Data in Brief 57 (2024), p. 110979

  24. [32]

    Age-Related Eye Disease Study Research Group et al. “The Age-Related Eye Disease Study sys- tem for classifying age-related macular degeneration from stereoscopic color fundus photographs: the Age-Related Eye Disease Study Report Number 6”. In: American journal of ophthalmolog...

  25. [33]

    The Age-Related Eye Disease Study (AREDS) system for classifying cataracts from photographs: AREDS report no. 4

    Age-Related Eye Disease Study Research Group et al. “The Age-Related Eye Disease Study (AREDS) system for classifying cataracts from photographs: AREDS report no. 4”. In:American journal of ophthalmology131.2 (2001), pp. 167–175

  26. [34]

    DREAM: diabetic retinopa- thy analysis using machine learning

    Sohini Roychowdhury, Dara D Koozekanani, and Keshab K Parhi. “DREAM: diabetic retinopa- thy analysis using machine learning”. In: IEEE journal of biomedical and health informatics18.5 (2013), pp. 1717–1728

  27. [35]

    Imagenet classification with deep convolutional neural networks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. “Imagenet classification with deep convolutional neural networks”. In:Advances in neural information processing systems25 (2012)

  28. [36]

    A survey on deep learning in medical image analysis

    Geert Litjens et al. “A survey on deep learning in medical image analysis”. In: Medical image analysis 42 (2017), pp. 60–88

  29. [37]

    A review on medical image applications based on deep learning techniques

    Ali H Abdulwahhab et al. “A review on medical image applications based on deep learning techniques”. In: Journal of Image and Graphics12.3 (2024), pp. 215–227

  30. [38]

    Analysis of deep learning techniques for prediction of eye diseases: A systematic review

    Akanksha Bali and Vibhakar Mansotra. “Analysis of deep learning techniques for prediction of eye diseases: A systematic review”. In: Archives of Computational Methods in Engineering31.1 (2024), pp. 487–520

  31. [39]

    A study of CNN and transfer learning in medical imaging: Advan- tages, challenges, future scope

    Ahmad Waleed Salehi et al. “A study of CNN and transfer learning in medical imaging: Advan- tages, challenges, future scope”. In: Sustainability 15.7 (2023), p. 5930

  32. [40]

    Review of deep learning: concepts, CNN architectures, challenges, appli- cations, future directions

    Laith Alzubaidi et al. “Review of deep learning: concepts, CNN architectures, challenges, appli- cations, future directions”. In: Journal of big Data8 (2021), pp. 1–74

  33. [41]

    The application of artificial intelligence in glaucoma diagnosis and predic- tion

    Linyu Zhang et al. “The application of artificial intelligence in glaucoma diagnosis and predic- tion”. In: Frontiers in cell and developmental biology11 (2023), p. 1173094

  34. [42]

    Artificial intelligence in myopia: current and future trends

    Li Lian Foo et al. “Artificial intelligence in myopia: current and future trends”. In: Current Opinion in Ophthalmology32.5 (2021), pp. 413–424. 16

  35. [43]

    Attention is all you need

    Ashish Vaswani et al. “Attention is all you need”. In: Advances in neural information processing systems 30 (2017)

  36. [44]

    Multi-label classification of retinal disease via a novel vision transformer model

    Dong Wang, Jian Lian, and Wanzhen Jiao. “Multi-label classification of retinal disease via a novel vision transformer model”. In: Frontiers in Neuroscience17 (2024), p. 1290803

  37. [45]

    Retinitis pigmentosa

    Dyonne T Hartong, Eliot L Berson, and Thaddeus P Dryja. “Retinitis pigmentosa”. In: The Lancet 368.9549 (2006), pp. 1795–1809

  38. [46]

    Pathology and pathogenesis of retinal detachment

    NG Ghazi and WR Green. “Pathology and pathogenesis of retinal detachment”. In: Eye 16.4 (2002), pp. 411–421

  39. [47]

    “Myopia”

    Ian G Morgan, Kyoko Ohno-Matsui, and Seang-Mei Saw. “Myopia”. In: The Lancet 379.9827 (2012), pp. 1739–1748

  40. [48]

    Gene and cell therapy for age-related macular degeneration: A review

    Jos´ e Trincao-Marques et al. “Gene and cell therapy for age-related macular degeneration: A review”. In: Survey of Ophthalmology(2024)

  41. [49]

    A review on glaucoma: causes, symptoms, pathogenesis & treatment

    Mahendra Kumar Sahu. “A review on glaucoma: causes, symptoms, pathogenesis & treatment”. In: Journal of Clinical Research and Ophthalmology11.1 (2024), pp. 001–004

  42. [50]

    Optic disc edema

    Gregory P Van Stavern. “Optic disc edema”. In: Seminars in neurology. Vol. 27. 03. Copyright© 2007 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New . . . 2007, pp. 233–243

  43. [51]

    A review on diabetic retinopathy

    Vijender Kour et al. “A review on diabetic retinopathy”. In: Current Diabetes Reviews 20.6 (2024), pp. 74–88

  44. [52]

    Central serous chorioretinopathy: a review

    Adrian T Fung, Yi Yang, and Andrew W Kam. “Central serous chorioretinopathy: a review”. In: Clinical & experimental ophthalmology51.3 (2023), pp. 243–270

  45. [53]

    Feature pyramid networks for object detection

    Tsung-Yi Lin et al. “Feature pyramid networks for object detection”. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2017, pp. 2117–2125

  46. [54]

    Squeeze-and-excitation networks

    Jie Hu, Li Shen, and Gang Sun. “Squeeze-and-excitation networks”. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2018, pp. 7132–7141. 17

Pith tools

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