pith. sign in

arxiv: 2605.17236 · v1 · pith:SIK7S6L4new · submitted 2026-05-17 · 💻 cs.CV · cs.AI

Systematic Evaluation of Vision Transformers for Automated Cervical Cancer Classification: Optimization, Statistical Validation, and Clinical Interpretability

Nisreen Albzour , Sarah S. Lam This is my paper

Pith reviewed 2026-05-20 15:07 UTC · model grok-4.3

classification 💻 cs.CV cs.AI
keywords Vision TransformerCervical Cancer ClassificationPap SmearMedical Image AnalysisGrad-CAMInterpretabilityHerlev Dataset
0
0 comments X

The pith

Vision Transformers reach 95 percent accuracy on cervical cell images while focusing on the same features doctors examine.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper evaluates whether Vision Transformers can automate Pap smear analysis for cervical cancer by optimizing a compact model on a dataset of 917 images. Through testing image flips, class balancing, and hyperparameters, the authors report 94.9 to 95.2 percent cross-validation accuracy. They then apply gradient-based attention mapping to verify that the model highlights nuclear regions, cell boundaries, and chromatin patterns that match cytopathology standards. This combination of accuracy and visual explanations addresses two barriers that have limited earlier neural networks in medical screening. If the result holds, it points toward transformer-based tools that could support consistent, transparent decisions in cervical cancer programs.

Core claim

The authors show that a systematically tuned Vision Transformer (ViT-Tiny) classifies normal versus abnormal cervical cells from the Herlev dataset at 94.9-95.2 percent accuracy when random horizontal flipping and class weights of 0.7 and 1.3 are applied, and that Grad-CAM attention maps align with clinically relevant morphological features including nuclear areas, cell edges, and chromatin texture.

What carries the argument

The lightweight Vision Transformer (ViT-Tiny) architecture, tuned through augmentation choices, class weighting, and hyperparameter search, with Grad-CAM used to produce attention visualizations that correspond to cytopathological criteria.

If this is right

  • Automated tools could reduce the time and observer-to-observer differences that affect manual Pap smear review.
  • Attention maps that match nuclear and chromatin features give a concrete basis for using the outputs in clinical decision support.
  • The identified combination of flipping and class weighting supplies a direct starting point for applying similar transformers to other cell-image tasks.
  • Meeting both accuracy and transparency thresholds removes a common obstacle to integrating AI into cervical screening workflows.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same optimization steps could be applied to multi-class grading of abnormalities to guide more precise follow-up decisions.
  • Pairing the model with images from multiple centers would reveal whether the reported accuracy survives domain shifts in staining or scanner type.
  • The built-in interpretability opens a path to systems that flag only the cases where attention deviates from expected clinical regions for human review.

Load-bearing premise

The Herlev collection of 917 images is representative enough that the optimized model will perform at similar accuracy on new images from varied clinical sources and equipment.

What would settle it

Testing the final model on an independent set of cervical smear images collected at a different site or population and observing accuracy fall below 85 percent.

read the original abstract

Manual Pap smear analysis for cervical cancer screening is limited by inter-observer variability, time constraints, and restricted expert availability. Although convolutional neural networks (CNNs) have automated cervical cell classification, they remain limited in modeling long-range spatial dependencies and often lack clinical interpretability. In this study, Vision Transformer (ViT) architectures were systematically optimized to enhance automated cervical cancer screening, which resulted in improved interpretability. The Herlev dataset (917 images: 242 normal, 675 abnormal) was utilized to optimize ViT-Tiny, a lightweight Vision Transformer architecture designed for reduced computational complexity, through a comprehensive evaluation of augmentation strategies, class weighting, and hyperparameters. The optimal configuration achieved 94.9%-95.2% cross-validation accuracy, in which random horizontal flipping and class weighting (0.7 x 1.3) were identified as most effective. Gradient-weighted Class Activation Mapping (Grad-CAM) analysis confirmed that model attention corresponded to clinically relevant morphological features, which include nuclear regions, cell boundaries, and chromatin texture, which align with cytopathological criteria. These findings indicate that Vision Transformers can deliver accurate and interpretable decision support for cervical cancer screening, which fulfills both clinical performance and transparency requirements essential for medical AI deployment.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript evaluates Vision Transformer (ViT-Tiny) architectures for automated classification of cervical cancer cells on the Herlev dataset (917 images: 242 normal, 675 abnormal). It optimizes augmentation strategies, class weighting (0.7 × 1.3), and hyperparameters via 5-fold cross-validation, reporting 94.9–95.2% accuracy. Grad-CAM visualizations are presented to show that attention aligns with clinically relevant features including nuclear regions, cell boundaries, and chromatin texture. The authors conclude that optimized ViTs can provide accurate, interpretable decision support suitable for clinical deployment in cervical cancer screening.

Significance. If the results hold under external scrutiny, the work would contribute to demonstrating that lightweight Vision Transformers can model long-range spatial dependencies in cytology images while offering Grad-CAM-based interpretability that maps to cytopathological criteria. The emphasis on a computationally efficient ViT-Tiny variant and systematic hyperparameter search could guide practical deployment in resource-constrained screening programs. The explicit linkage of model attention to morphological features addresses a key barrier to clinical adoption of medical AI.

major comments (2)
  1. [Abstract] Abstract: The central claim that the approach 'fulfills both clinical performance and transparency requirements essential for medical AI deployment' is load-bearing yet rests solely on internal 5-fold CV results from a single-source dataset. No external validation on independent cohorts, different scanners, staining protocols, or multi-center collections is reported, leaving domain-shift risks (common in Pap-smear imaging) unquantified.
  2. [Abstract] Abstract: The title references 'Statistical Validation,' but the provided description supplies only aggregate accuracy ranges without confidence intervals, p-values for baseline comparisons, error analysis stratified by class, or explicit handling of the 242:675 class imbalance beyond the reported weighting factor.
minor comments (1)
  1. [Abstract] The abstract would benefit from a concise statement of the exact baseline CNN performance and the precise statistical tests used to support the reported accuracy figures.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. These observations help clarify the scope of our evaluation and the strength of our claims. We address each major comment point by point below and indicate the revisions we will make to the abstract and main text.

read point-by-point responses

  1. Referee: [Abstract] The central claim that the approach 'fulfills both clinical performance and transparency requirements essential for medical AI deployment' is load-bearing yet rests solely on internal 5-fold CV results from a single-source dataset. No external validation on independent cohorts, different scanners, staining protocols, or multi-center collections is reported, leaving domain-shift risks (common in Pap-smear imaging) unquantified.

    Authors: We agree that external validation on independent cohorts is necessary to quantify domain-shift risks and to support broad claims of clinical suitability. Our study is a systematic benchmark evaluation and hyperparameter optimization performed on the publicly available Herlev dataset using 5-fold cross-validation. We will revise the abstract to qualify the central claim, stating that the optimized ViT-Tiny delivers high accuracy and Grad-CAM interpretability aligned with cytopathological features on this dataset, thereby providing evidence toward meeting clinical performance and transparency needs while explicitly noting the requirement for multi-center validation. A new limitations subsection will discuss generalizability and domain-shift considerations. revision: yes

  2. Referee: [Abstract] The title references 'Statistical Validation,' but the provided description supplies only aggregate accuracy ranges without confidence intervals, p-values for baseline comparisons, error analysis stratified by class, or explicit handling of the 242:675 class imbalance beyond the reported weighting factor.

    Authors: The full manuscript presents 5-fold cross-validation results with the reported class weighting (0.7 × 1.3). To better substantiate the 'Statistical Validation' component of the title, we will add 95% confidence intervals for all accuracy metrics, include statistical significance testing (e.g., paired t-tests or McNemar's test) against baseline CNN models, report per-class error rates and confusion matrices, and expand the methods section with a detailed rationale for the weighting scheme and its effect on the imbalanced 242:675 distribution. These changes will be reflected in an updated abstract if space permits. revision: yes

Circularity Check

0 steps flagged

Empirical ML evaluation with cross-validation exhibits no circular derivation

full rationale

The paper conducts a standard empirical optimization and evaluation of Vision Transformer architectures on the fixed Herlev dataset of 917 images. It reports cross-validation accuracies after testing augmentation strategies, class weighting, and hyperparameters, followed by post-hoc Grad-CAM visualization. No derivation chain, first-principles result, or prediction is claimed that reduces by construction to the fitted parameters or input data splits. Hyperparameter selection is data-dependent in the usual ML sense but does not rename a fit as an independent prediction. No self-citation load-bearing steps, uniqueness theorems, or ansatzes imported from prior author work appear in the provided text. The work is self-contained as an experimental benchmark study.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard supervised learning assumptions plus the representativeness of a single public dataset; no new entities postulated.

free parameters (1)
  • class weighting factors = 0.7 and 1.3
    0.7 x 1.3 factors chosen to address imbalance between normal and abnormal classes.
axioms (1)
  • domain assumption Labels in the Herlev dataset accurately distinguish normal from abnormal cervical cells.
    Required for supervised training and reported accuracy metrics.

pith-pipeline@v0.9.0 · 5757 in / 1325 out tokens · 68540 ms · 2026-05-20T15:07:24.721298+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

60 extracted references · 60 canonical work pages · 1 internal anchor

  1. [1]

    Introduction Cervical cancer remains a significant global health challenge, in which over 600,000 new cases and 340,000 deaths are reported annually worldwide [1]. The primary strategy for reducing mortality is early detection through Pap smear screening, which enables identification of precancerous lesions before they progress to invasive cancer. However...

    work page

  2. [2]

    Related Literature The application of artificial intelligence techniques in cervical cancer detection and classification has advanced significantly over the past decade, which led to the development of increasingly accurate and automated diagnostic methods. As illustrated in Figure 1, these techniques can be broadly categorized into machine learning and d...

    work page

  3. [3]

    Methodology Figure 2 provides a comprehensive overview of the entire methodology employed in this study. Each stage plays a crucial role that ensures the effectiveness and robustness of the classification framework, which underscores the systematic approach adopted to tackle the challenges in cervical cancer diagnosis. Figure 2: Overview of the Methodolog...

    work page

  4. [4]

    Organizin g results in parallel with methodology enables direct tracing from design choices to performance

    Results and Discussion This section reports experimental outcomes following the same five -stage workflow described in Section 3 and Figure 2: (1) Dataset Preparation, (2) Data Augmentation, (3) Class Weighting, (4) ViT Model Training, and (5) Grad -CAM Interpretability. Organizin g results in parallel with methodology enables direct tracing from design c...

    work page

  5. [5]

    Conclusion This study demonstrates the potential of Vision Transformers as accurate and interpretable tools for automated cervical cancer screening. By restructuring Pap smear classification into a binary screening task, the experiments were aligned with clinical pra ctice and demonstrated that optimized ViT models can achieve ~95% accuracy with statistic...

    work page

  6. [6]

    Nayar and D

    R. Nayar and D. C. Wilbur, Eds., The Bethesda System for Reporting Cervical Cytology: Definitions, Criteria, and Explanatory Notes. Cham: Springer International Publishing, 2015. doi: 10.1007/978-3-319-11074-5

    work page doi:10.1007/978-3-319-11074-5 2015

  7. [7]

    Deep learning,

    Y . LeCun, Y . Bengio, and G. Hinton, “Deep learning, ” nature, vol. 521, no. 7553, pp. 436–444, 2015

    work page 2015

  8. [8]

    Segmentation and classification techniques for Pap smear images in detecting cervical cancer: a systematic review,

    B. Z. Wubineh, A. Rusiecki, and K. Halawa, “Segmentation and classification techniques for Pap smear images in detecting cervical cancer: a systematic review, ” IEEE Access, 2024

    work page 2024

  9. [9]

    U-Net: Convolutional Networks for Biomedical Image Segmentation

    O. Ronneberger, P . Fischer, and T. Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation, ” May 18, 2015, arXiv: arXiv:1505.04597. doi: 10.48550/arXiv.1505.04597

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1505.04597 2015

  10. [10]

    Khoshgoftaar

    C. Shorten and T. M. Khoshgoftaar, “A survey on Image Data Augmentation for Deep Learning, ” J. Big Data, vol. 6, no. 1, p. 60, Dec. 2019, doi: 10.1186/s40537-019-0197-0

    work page doi:10.1186/s40537-019-0197-0 2019

  11. [11]

    Toward driverless AI: Automating leukemia detection and classification using hyperautomation, a case study,

    H. Al-Zoubi and N. Al-Bzoor, “Toward driverless AI: Automating leukemia detection and classification using hyperautomation, a case study, ” 2022

    work page 2022

  12. [12]

    Guest Editorial Deep Learning in Medical Imaging: Overview and Future Promise of an Exciting New Technique,

    H. Greenspan, B. Van Ginneken, and R. M. Summers, “Guest Editorial Deep Learning in Medical Imaging: Overview and Future Promise of an Exciting New Technique, ” IEEE Trans. Med. Imaging, vol. 35, no. 5, pp. 1153–1159, May 2016, doi: 10.1109/TMI.2016.2553401

    work page doi:10.1109/tmi.2016.2553401 2016

  13. [13]

    van der Laak, Bram van Ginneken, and Clara I

    G. Litjens et al., “A survey on deep learning in medical image analysis, ” Med. Image Anal., vol. 42, pp. 60–88, Dec. 2017, doi: 10.1016/j.media.2017.07.005

    work page doi:10.1016/j.media.2017.07.005 2017

  14. [14]

    Combining machine learning and deep learning approaches to detect cervical cancer in cytology images,

    E. L. P . da Silva, “Combining machine learning and deep learning approaches to detect cervical cancer in cytology images, ” 2021

    work page 2021

  15. [15]

    Automatic model for cervical cancer screening based on convolutional neural network: a retrospective, multicohort, multicenter study,

    X. Tan et al., “Automatic model for cervical cancer screening based on convolutional neural network: a retrospective, multicohort, multicenter study, ” Cancer Cell Int., vol. 21, no. 1, p. 35, Dec. 2021, doi: 10.1186/s12935-020-01742-6

    work page doi:10.1186/s12935-020-01742-6 2021

  16. [16]

    Axiomatic attribution for deep networks,

    M. Sundararajan, A. Taly, and Q. Yan, “Axiomatic attribution for deep networks, ” presented at the International conference on machine learning, PMLR, 2017, pp. 3319– 3328

    work page 2017

  17. [17]

    Explainable artificial intelligence: Understanding, visualizing and interpreting deep learning models,

    W. Samek, T. Wiegand, and K.-R. Müller, “Explainable artificial intelligence: Understanding, visualizing and interpreting deep learning models, ” ArXiv Prepr. ArXiv170808296, 2017

    work page 2017

  18. [18]

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

    A. Dosovitskiy et al., “An image is worth 16x16 words: Transformers for image recognition at scale, ” ArXiv Prepr. ArXiv201011929, 2020

    work page 2020

  19. [19]

    Attention is all you need,

    A. Vaswani et al., “Attention is all you need, ” Adv. Neural Inf. Process. Syst., vol. 30, 2017

    work page 2017

  20. [20]

    A survey on vision transformer,

    K. Han et al., “A survey on vision transformer, ” IEEE Trans. Pattern Anal. Mach. Intell., vol. 45, no. 1, pp. 87–110, 2022

    work page 2022

  21. [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, ” presented at the Proceedings of the IEEE international conference on computer vision, 2017, pp. 618– 626

    work page 2017

  22. [22]

    An effective CNN and Transformer complementary network for medical image segmentation,

    F . Yuan, Z. Zhang, and Z. Fang, “An effective CNN and Transformer complementary network for medical image segmentation, ” Pattern Recognit., vol. 136, p. 109228, 2023

    work page 2023

  23. [23]

    UCTNet: Uncertainty-guided CNN-Transformer hybrid networks for medical image segmentation,

    X. Guo, X. Lin, X. Yang, L. Yu, K.-T. Cheng, and Z. Yan, “UCTNet: Uncertainty-guided CNN-Transformer hybrid networks for medical image segmentation, ” Pattern Recognit., vol. 152, p. 110491, 2024

    work page 2024

  24. [24]

    A hybrid linear iterative clustering and Bayes classification- based GrabCut segmentation scheme for dynamic detection of cervical cancer,

    A. D. Magaraja et al., “A hybrid linear iterative clustering and Bayes classification- based GrabCut segmentation scheme for dynamic detection of cervical cancer, ” Appl. Sci., vol. 12, no. 20, p. 10522, 2022

    work page 2022

  25. [25]

    Knn and Fuzzy Logic for Segmentation and Classification of Cervical Cancer from Pap‑Smear Images,

    S. Ben Chaabane, A. Bushnag, W. Massoudi, A. Oyouni, and S. Abuzneid, “Knn and Fuzzy Logic for Segmentation and Classification of Cervical Cancer from Pap‑Smear Images, ” Sep. 12, 2024, Social Science Research Network, Rochester, NY: 4942634. doi: 10.2139/ssrn.4942634

    work page doi:10.2139/ssrn.4942634 2024

  26. [26]

    Machine Learning Assisted Cervical Cancer Detection,

    M. Mehmood, M. Rizwan, M. Gregus Ml, and S. Abbas, “Machine Learning Assisted Cervical Cancer Detection, ” Front. Public Health, vol. 9, p. 788376, Dec. 2021, doi: 10.3389/fpubh.2021.788376

    work page doi:10.3389/fpubh.2021.788376 2021

  27. [27]

    Nominated Texture Based Cervical Cancer Classification,

    E. J. Mariarputham and A. Stephen, “Nominated Texture Based Cervical Cancer Classification, ” Comput. Math. Methods Med., vol. 2015, pp. 1–10, 2015, doi: 10.1155/2015/586928

    work page doi:10.1155/2015/586928 2015

  28. [28]

    A quantum hybrid PSO combined with fuzzy k-NN approach to feature selection and cell classification in cervical cancer detection,

    A. M. Iliyasu and C. Fatichah, “A quantum hybrid PSO combined with fuzzy k-NN approach to feature selection and cell classification in cervical cancer detection, ” Sensors, vol. 17, no. 12, p. 2935, 2017

    work page 2017

  29. [29]

    Predicting Cervical Cancer Cases Resulting in Biopsies Using Machine Learning Techniques,

    Tenali Pranuthi, “Predicting Cervical Cancer Cases Resulting in Biopsies Using Machine Learning Techniques, ” Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol., pp. 28–37, Nov. 2021, doi: 10.32628/CSEIT217616

    work page doi:10.32628/cseit217616 2021

  30. [30]

    Predicting Cervical Cancer using Machine Learning Methods,

    R. Alsmariy, G. Healy, and H. Abdelhafez, “Predicting Cervical Cancer using Machine Learning Methods, ” Int. J. Adv. Comput. Sci. Appl., vol. 11, no. 7, 2020, doi: 10.14569/IJACSA.2020.0110723

    work page doi:10.14569/ijacsa.2020.0110723 2020

  31. [31]

    SKCV: Stratified K-fold cross-validation on ML classifiers for predicting cervical cancer,

    S. Prusty, S. Patnaik, and S. K. Dash, “SKCV: Stratified K-fold cross-validation on ML classifiers for predicting cervical cancer, ” Front. Nanotechnol., vol. 4, p. 972421, 2022

    work page 2022

  32. [32]

    Pap Smear Images Classification for Early Detection of Cervical Cancer,

    A. HassanMbaga and P . ZhiJun, “Pap Smear Images Classification for Early Detection of Cervical Cancer, ” Int. J. Comput. Appl., vol. 118, no. 7, pp. 10–16, May 2015, doi: 10.5120/20756-3159

    work page doi:10.5120/20756-3159 2015

  33. [33]

    Early prediction of cervical cancer using machine learning techniques,

    M. S. Al-Batah, M. Alzyoud, R. Alazaidah, M. Toubat, H. Alzoubi, and A. Olaiyat, “Early prediction of cervical cancer using machine learning techniques, ” Jordanian J. Comput. Inf. Technol., vol. 8, no. 4, 2022

    work page 2022

  34. [34]

    A Novel Approach to Cervical Cancer Detection Using Hybrid Stacked Ensemble Models and Feature Selection,

    P . D. Nandanwar and Dr. S. B. Dhonde, “A Novel Approach to Cervical Cancer Detection Using Hybrid Stacked Ensemble Models and Feature Selection, ” Int. J. Electr. Electron. Res., vol. 11, no. 2, pp. 582–589, Jun. 2023, doi: 10.37391/ijeer.110246

    work page doi:10.37391/ijeer.110246 2023

  35. [35]

    Cervical Cancer Classification,

    D. K, “Cervical Cancer Classification, ” Int. J. Emerg. Trends Eng. Res., vol. 8, no. 3, pp. 804–807, Mar. 2020, doi: 10.30534/ijeter/2020/32832020

    work page doi:10.30534/ijeter/2020/32832020 2020

  36. [36]

    Cervical image classification based on image segmentation preprocessing and a CapsNet network model,

    X. Zhang and S. Zhao, “Cervical image classification based on image segmentation preprocessing and a CapsNet network model, ” Int. J. Imaging Syst. Technol., vol. 29, no. 1, pp. 19–28, Mar. 2019, doi: 10.1002/ima.22291

    work page doi:10.1002/ima.22291 2019

  37. [37]

    Detection of Cervical Cancer Cells in Whole Slide Images Using Deformable and Global Context Aware Faster RCNN- FPN,

    X. Li, Z. Xu, X. Shen, Y . Zhou, B. Xiao, and T.-Q. Li, “Detection of Cervical Cancer Cells in Whole Slide Images Using Deformable and Global Context Aware Faster RCNN- FPN, ” Curr. Oncol., vol. 28, no. 5, pp. 3585–3601, Sep. 2021, doi: 10.3390/curroncol28050307

    work page doi:10.3390/curroncol28050307 2021

  38. [38]

    Cervical Transformation Zone Segmentation and Classification based on Improved Inception-ResNet-V2 Using Colposcopy Images,

    S. Dash, P . K. Sethy, and S. K. Behera, “Cervical Transformation Zone Segmentation and Classification based on Improved Inception-ResNet-V2 Using Colposcopy Images, ” Cancer Inform., vol. 22, Jan. 2023, doi: 10.1177/11769351231161477

    work page doi:10.1177/11769351231161477 2023

  39. [39]

    Multi-class Cervical Cancer Classification using Transfer Learning-based Optimized SE-ResNet152 model in Pap Smear Whole Slide Images,

    K. Prasad Battula and B. Sai Chandana, “Multi-class Cervical Cancer Classification using Transfer Learning-based Optimized SE-ResNet152 model in Pap Smear Whole Slide Images, ” Int. J. Electr. Comput. Eng. Syst., vol. 14, no. 6, pp. 623–623, Jul. 2023, doi: 10.32985/ijeces.14.6.1

    work page doi:10.32985/ijeces.14.6.1 2023

  40. [40]

    Privacy Preserved Cervical Cancer Detection Using Convolutional Neural Networks Applied to Pap Smear Images,

    S. Alsubai et al., “Privacy Preserved Cervical Cancer Detection Using Convolutional Neural Networks Applied to Pap Smear Images, ” Comput. Math. Methods Med., vol. 2023, no. 1, Jan. 2023, doi: 10.1155/2023/9676206

    work page doi:10.1155/2023/9676206 2023

  41. [41]

    Hybrid Loss-Constrained Lightweight Convolutional Neural Networks for Cervical Cell Classification,

    W. Chen, W. Shen, L. Gao, and X. Li, “Hybrid Loss-Constrained Lightweight Convolutional Neural Networks for Cervical Cell Classification, ” Sensors, vol. 22, no. 9, p. 3272, Apr. 2022, doi: 10.3390/s22093272

    work page doi:10.3390/s22093272 2022

  42. [42]

    Automated Cervical Cancer Screening Using Single-Cell Segmentation and Deep Learning: Enhanced Performance with Liquid-Based Cytology,

    M. Rodríguez, C. Córdova, I. Benjumeda, and S. San Martín, “Automated Cervical Cancer Screening Using Single-Cell Segmentation and Deep Learning: Enhanced Performance with Liquid-Based Cytology, ” Computation, vol. 12, no. 12, p. 232, Nov. 2024, doi: 10.3390/computation12120232

    work page doi:10.3390/computation12120232 2024

  43. [43]

    AF-SENet: Classification of Cancer in Cervical Tissue Pathological Images Based on Fusing Deep Convolution Features,

    P . Huang, X. Tan, C. Chen, X. Lv, and Y . Li, “AF-SENet: Classification of Cancer in Cervical Tissue Pathological Images Based on Fusing Deep Convolution Features, ” Sensors, vol. 21, no. 1, p. 122, Dec. 2020, doi: 10.3390/s21010122

    work page doi:10.3390/s21010122 2020

  44. [44]

    Segmentation and Classification of Cervical Cells Using Deep Learning,

    Kurnianingsih et al., “Segmentation and Classification of Cervical Cells Using Deep Learning, ” IEEE Access, vol. 7, pp. 116925–116941, 2019, doi: 10.1109/access.2019.2936017

    work page doi:10.1109/access.2019.2936017 2019

  45. [45]

    Deep Learning based Cervical Cancer Classification and Segmentation from Pap Smears Images using an EfficientNet,

    K. P . Battula and B. S. Chandana, “Deep Learning based Cervical Cancer Classification and Segmentation from Pap Smears Images using an EfficientNet, ” Int. J. Adv. Comput. Sci. Appl., vol. 13, no. 9, 2022, doi: 10.14569/IJACSA.2022.01309104

    work page doi:10.14569/ijacsa.2022.01309104 2022

  46. [46]

    Classification of cervical cancer using Dense CapsNet with Seg-UNet and denoising autoencoders,

    H. Yang, W. Aydi, N. Innab, M. E. Ghoneim, and M. Ferrara, “Classification of cervical cancer using Dense CapsNet with Seg-UNet and denoising autoencoders, ” Sci. Rep., vol. 14, no. 1, Dec. 2024, doi: 10.1038/s41598-024-82489-2

    work page doi:10.1038/s41598-024-82489-2 2024

  47. [47]

    Multi-Class Segmentation with Deep Learning based Pap Smear Image Analysis for Cervical Cancer Detection and Classification Model,

    C. Suguna, Dr. S. P . Balamurugan, “Multi-Class Segmentation with Deep Learning based Pap Smear Image Analysis for Cervical Cancer Detection and Classification Model, ” Tuijin JishuJournal Propuls. Technol., vol. 44, no. 3, pp. 4475–4487, Nov. 2023, doi: 10.52783/tjjpt.v44.i3.2453

    work page doi:10.52783/tjjpt.v44.i3.2453 2023

  48. [48]

    Segmentation and Classification of Pap Smear Images for Cervical Cancer Detection Using Deep Learning,

    N. Albzour and S. S. Lam, “Segmentation and Classification of Pap Smear Images for Cervical Cancer Detection Using Deep Learning, ” arXiv preprint arXiv:2508.17728, 2025. doi: 10.48550/arXiv.2508.17728

    work page doi:10.48550/arxiv.2508.17728 2025

  49. [49]

    Nucleus segmentation and classification using residual SE-UNet and feature concatenation approach incervical cytopathology cell images,

    G. J. Chowdary, S. G, P . M, and P . Yogarajah, “Nucleus segmentation and classification using residual SE-UNet and feature concatenation approach incervical cytopathology cell images, ” Technol. Cancer Res. Treat., vol. 22, Jan. 2023, doi: 10.1177/15330338221134833

    work page doi:10.1177/15330338221134833 2023

  50. [50]

    Cell classification framework using U-Net: convolutional networks for cervix cell segmentation,

    T. Ermis, E. Sener, and M. Elitas, “Cell classification framework using U-Net: convolutional networks for cervix cell segmentation, ” in Emerging Topics in Artificial Intelligence (ETAI) 2023, G. Volpe, J. B. Pereira, D. Brunner, and A. Ozcan, Eds., San Diego, United States: SPIE, Sep. 2023, p. 64. doi: 10.1117/12.2677423

    work page doi:10.1117/12.2677423 2023

  51. [51]

    Encoder-Weighted W-Net for Unsupervised Segmentation of Cervix Region in Colposcopy Images,

    J. Park, H. Yang, H.-J. Roh, W. Jung, and G.-J. Jang, “Encoder-Weighted W-Net for Unsupervised Segmentation of Cervix Region in Colposcopy Images, ” Cancers, vol. 14, no. 14, p. 3400, Jul. 2022, doi: 10.3390/cancers14143400

    work page doi:10.3390/cancers14143400 2022

  52. [52]

    DeepCyto: a hybrid framework for cervical cancer classification by using deep feature fusion of cytology images,

    S. Shinde, M. Kalbhor, and P . Wajire, “DeepCyto: a hybrid framework for cervical cancer classification by using deep feature fusion of cytology images, ” Math. Biosci. Eng., vol. 19, no. 7, pp. 6415–6434, 2022, doi: 10.3934/mbe.2022301

    work page doi:10.3934/mbe.2022301 2022

  53. [53]

    DeepCerviCancer - Deep Learning-Based Cervical Image Classification using Colposcopy and Cytology Images,

    M. Kalbhor, S. Shinde, S. Lahade, and T. Choudhury, “DeepCerviCancer - Deep Learning-Based Cervical Image Classification using Colposcopy and Cytology Images, ” EAI Endorsed Trans. Pervasive Health Technol., vol. 9, Oct. 2023, doi: 10.4108/eetpht.9.3473

    work page doi:10.4108/eetpht.9.3473 2023

  54. [54]

    Hybridization of Deep Learning Pre-Trained Models with Machine Learning Classifiers and Fuzzy Min–Max Neural Network for Cervical Cancer Diagnosis,

    M. Kalbhor, S. Shinde, D. E. Popescu, and D. J. Hemanth, “Hybridization of Deep Learning Pre-Trained Models with Machine Learning Classifiers and Fuzzy Min–Max Neural Network for Cervical Cancer Diagnosis, ” Diagnostics, vol. 13, no. 7, p. 1363, Apr. 2023, doi: 10.3390/diagnostics13071363

    work page doi:10.3390/diagnostics13071363 2023

  55. [55]

    Radiomics Diagnostic Tool Based on Deep Learning for Colposcopy Image Classification,

    Y . Jiménez Gaona et al., “Radiomics Diagnostic Tool Based on Deep Learning for Colposcopy Image Classification, ” Diagnostics, vol. 12, no. 7, p. 1694, Jul. 2022, doi: 10.3390/diagnostics12071694

    work page doi:10.3390/diagnostics12071694 2022

  56. [56]

    Automatic cervical cancer classification using adaptive vision transformer encoder with CNN for medical application,

    G. Nirmala, P . P . Nayudu, A. R. Kumar, and R. Sagar, “Automatic cervical cancer classification using adaptive vision transformer encoder with CNN for medical application, ” Pattern Recognit., vol. 160, p. 111201, Apr. 2025, doi: 10.1016/j.patcog.2024.111201

    work page doi:10.1016/j.patcog.2024.111201 2025

  57. [57]

    Multi-objective optimization of ViT architecture for efficient brain tumor classification,

    E. Şahin, D. Özdemir, and H. Temurtaş, “Multi-objective optimization of ViT architecture for efficient brain tumor classification, ” Biomed. Signal Process. Control, vol. 91, p. 105938, 2024

    work page 2024

  58. [58]

    Sstrans-net: Smart swin transformer network for medical image segmentation,

    L. Fu, Y . Chen, W. Ji, and F . Yang, “Sstrans-net: Smart swin transformer network for medical image segmentation, ” Biomed. Signal Process. Control, vol. 91, p. 106071, 2024

    work page 2024

  59. [59]

    Hybrid CNN-Transformer model for medical image segmentation with pyramid convolution and multi-layer perceptron,

    X. Liu, Y . Hu, and J. Chen, “Hybrid CNN-Transformer model for medical image segmentation with pyramid convolution and multi-layer perceptron, ” Biomed. Signal Process. Control, vol. 86, p. 105331, 2023

    work page 2023

  60. [60]

    Predicting post-stroke activities of daily living: Enhancing Machine Learning with Feature Selection,

    N. Albzour, S. Agarwal, H. Althnaibat, and S. S. Lu, “Predicting post-stroke activities of daily living: Enhancing Machine Learning with Feature Selection, ” IISE Annual Conference Proceedings, pp. 1–6, 2025

    work page 2025