Pith. sign in

REVIEW 5 major objections 6 minor 43 references

Hybrid deep learning-based strategy for the hepatocellular carcinoma cancer grade classification of H&E stained liver histopathology images

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A hybrid convolutional network that fine-tunes the top layers of a pre-trained feature extractor and classifies through a deep fully connected stack reaches 100 percent accuracy on a public liver cancer slide dataset and 96.71 percent on…

desk verdict Patch-level splitting and test-set model selection undermine the headline results of this otherwise ordinary transfer-learning study; the 100% TCGA number is not evidence of slide-level generalization. read the letter →

arxiv 2412.03084 v2 pith:DDN2MVEZ submitted 2024-12-04 eess.IV cs.CVcs.LGq-bio.QM

Ajinkya Deshpande , Deep Gupta , Ankit Bhurane , Nisha Meshram , Sneha Singh , Petia Radeva This is my paper · ORCID
classification eess.IVcs.CVcs.LGq-bio.QM
keywords hepatocellularcarcinomahistopathologyimageclassificationhybriddeeplearningtransferfine-tuningpatch-basedH&Estainedslidesconvolutionalneuralnetwork
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 claims that a simple modification to transfer learning—freezing a pre-trained CNN's lower layers, fine-tuning its upper layers, and replacing its shallow classifier with a deeper fully connected stack—consistently improves patch-level classification of hepatocellular carcinoma in H&E stained liver slides. On the public development database the best hybrid reaches 100% accuracy, sensitivity, specificity, and F1-score with AUC 1.00, while on an independent four-class hospital database the best hybrid reaches 96.71% accuracy. The same strategy also pushes three backbones to 100% on a colon histopathology dataset. If the results hold under strict data separation, the work offers a low-effort recipe for boosting cancer grading without designing new architectures.

What carries the argument

The load-bearing mechanism is the hybrid model: a pre-trained convolutional feature extractor whose bottom layers stay frozen while its top layers are trainable, connected to a deep fully connected (ANN) classifier that gradually reduces feature dimensionality to the output label space. Compared with the base model—only the final layer retrained—the hybrid unfreezes a few top feature-extractor layers and adds several hidden layers. Training uses 5-fold stratified cross-validation with weighted random sampling for class balance and a cosine annealing warm-restart schedule; preprocessing includes patch extraction, a stain-vector color normalization method, and augmentation.

What would settle it

Run the same hybrid training and evaluation but assign all patches from each whole-slide image (or each patient) to either the train or the test fold, never both; if development-set accuracy drops well below 100% (or hospital accuracy below 96.71%), the original random-patch split was the source of the high scores. A complementary test: take the development-trained hybrid and evaluate it on slides from a different hospital with different staining, with no fine-tuning, and compare to the reported hospital accuracy.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that a hybrid architecture—a pre-trained CNN feature extractor with frozen bottom layers and selectively fine-tuned top layers, attached to a custom ANN classifier with gradual dimension reduction—outperforms the corresponding base pre-trained models across every dataset tested. The gain is consistent: the minimum accuracy increase is 0.76% on the development set and 3.29% on the hospital set, with one backbone achieving perfect scores on the three-class development task and another leading on the four-class hospital task. The paper interprets this as evidence that classifier depth and selective fine-tuning matter more than which backbone is chosen, and that no single architecture is universally best across datasets.

Load-bearing premise

The reported accuracies assume that randomly dividing image patches into training and test sets keeps the sets statistically independent, meaning no patch from the same whole-slide image appears in both sets; if that leaks, the model can memorize slide-specific staining and the numbers are inflated.

Editorial extensions

If this is right

  • The hybrid recipe can be applied to any pre-trained CNN and yields gains even when the base model already exceeds 99% accuracy, suggesting the bottleneck is classifier capacity, not feature extraction.
  • Near-perfect patch-level separation on the development set suggests automated triage of liver slides is feasible at patch level, though slide-level diagnosis remains untested.
  • The best backbone differs by dataset (ResNet50 on the development set, EfficientNetb3 on the hospital set), so practitioners should expect to select backbones per deployment site.
  • The same deep-classifier modification transfers to another histopathology domain (colon), indicating the strategy is not liver-specific.

Reading between the lines

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

  • Because the train/test split is made at the patch level rather than at the whole-slide or patient level, patches from the same slide can appear in both folds; a slide-holdout split would likely lower the reported numbers, and the 100% development-set result should be read with that caveat.
  • The accuracy gap between hybrid and base models may partly reflect the deeper classifier memorizing stain and tissue artifacts specific to each slide, not just tumor biology; testing on external slides from a different scanner or staining batch would separate these effects.
  • The stated comparison with prior work uses different patch sets and split protocols, so the headline gains over earlier methods are not a controlled comparison; a common benchmark split would be needed to verify superiority.
  • A practical extension would be to train the hybrid on patches but aggregate predictions at whole-slide level (e.g., majority vote), which is the clinically actionable unit and would test whether perfect patch accuracy translates to slide-level diagnosis.
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

5 major / 6 minor

Summary. This manuscript proposes a hybrid CNN-ANN architecture for hepatocellular carcinoma (HCC) grade classification from H&E-stained liver histopathology images. The method uses a pre-trained CNN (one of eight backbones) as a feature extractor, freezes its bottom layers, fine-tunes its top layers, and replaces the final classifier with a deep stack of fully connected layers. The authors evaluate the approach on the TCGA-LIHC dataset (3920 patches, three classes), a proprietary KMC dataset (3005 patches, four classes), and the LC25000 colon dataset (10,000 patches, two classes), using five-fold cross-validation and a 90/10 train/test split. They report that the hybrid model with ResNet50 achieves 100% accuracy on TCGA, that the hybrid model with EfficientNetb3 achieves 96.71% accuracy on KMC, and that hybrid models improve accuracy by roughly 2-5 percentage points over the corresponding base models.

Significance. If the reported results were obtained under a statistically valid evaluation protocol, the paper would offer a simple and reproducible recipe—selective fine-tuning of top convolutional layers plus a deep classifier—that improves several well-known backbones for histopathology patch classification. The systematic comparison across eight architectures and three datasets is a useful reference, and the public TCGA and LC25000 results would facilitate replication. However, the significance is currently undercut by severe methodological concerns about data leakage and test-set-based model selection, which make the absolute accuracy numbers and the claimed improvements unreliable as evidence of generalization to unseen slides.

major comments (5)
  1. [Section II.A, Section III.A, Table I] The train/test split is performed at the patch level rather than at the level of whole-slide images or patients. Section II.A describes tiling 491 TCGA WSIs into 1024×1024 patches, and Section III.A and Table I then divide the resulting 3920 patches into 3528 training and 392 test patches randomly. Because patches from the same WSI share staining, scanner, and tissue context, this protocol does not yield an independent test set, and the 100% TCGA and 96.71% KMC accuracies are likely inflated by the model memorizing slide-specific appearance. The authors must re-run the evaluation with a split that assigns all patches from a given WSI (or patient) to the same fold, and report metrics under that protocol.
  2. [Section II.A and Section III.A] The data augmentation described in Section II.A (random vertical and horizontal flips) is applied before the train/test split of Table I. Since the augmented counts (1220, 1340, 1360) are derived from the original extracted patches (813, 893, 680), flipped versions of the same original patch can appear in both training and test partitions, putting near-duplicates in the test set. The authors should split the data before any augmentation, or ensure that augmented copies of a given original patch are confined to the same partition.
  3. [Section III.C, Tables IV and VI] The best feature extractor is chosen by evaluating all eight candidate models on the same test set and then reporting the top performer (ResNet50 on TCGA, EfficientNetb3 on KMC) as the proposed result. This is a form of test-set fitting: the reported accuracy is the maximum over eight random variables and is therefore optimistically biased. The authors should select the architecture on a separate validation set (or via nested cross-validation) and report performance on a test set held out from all selection decisions.
  4. [Section II.C and Section III.C] The evaluation protocol is unclear: Section II.C describes five-fold cross-validation on the training data, but Section III.A and Table I introduce a separate 90/10 test split, and the results in Tables III-VIII are said to be 'averaged over five-folds.' It is not specified whether the test set is reused across folds or whether the reported numbers are cross-validation fold averages. This ambiguity makes the metrics difficult to interpret and must be resolved by a precise description of the training/validation/test workflow.
  5. [Section IV] The Limitations paragraph addresses only the high parameter count of the models. It does not mention the absence of slide-level splitting, the risk of augmentation leakage across the train/test boundary, or the optimistic bias from choosing the feature extractor on the test set. These are the most serious threats to the validity of the reported results and should be acknowledged and discussed.
minor comments (6)
  1. [Abstract and Section III.C] The claim of 'improvement in accuracy of 2% and 4%' is inconsistent with the detailed numbers (1.74 percentage points for TCGA in Section III.C.1 and 4.65 percentage points for KMC in Section III.C.2). Please use percentage points consistently.
  2. [Section II.A] The augmentation results in different multipliers per class (1220/813 ≈ 1.5, 1340/893 ≈ 1.5, 1360/680 = 2.0). The exact augmentation strategy should be described, particularly why the type-2 class received a larger number of augmented patches.
  3. [Figure 5] The 'offline system' and 'online system' are introduced without definitions; a one-sentence explanation would help the reader.
  4. [Section III.A] The sentence 'Cosine annealing warm restart [38] learning rate scheduler is used to select the learning rate over all epochs' should read '...to schedule the learning rate...'.
  5. [Table IX] The colon dataset is labeled 'LC25000' in Table IX but called 'COLON' elsewhere; unify the name.
  6. [Sections III.C.1 and III.C.2] The statement that the training curves indicate 'the absence of both bias and variance' is not established by the presented plots; consider replacing it with a more cautious description.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the hybrid-versus-base comparisons are empirical measurements, and no claimed result reduces to its own inputs by construction.

full rationale

The paper is an empirical deep-learning benchmark rather than a derivation. The central claim is that a hybrid model (frozen bottom CNN layers plus trainable top layers and a deep ANN classifier) outperforms a base model (frozen feature extractor with only the final layer retrained). This is supported by measured metrics on held-out test folds (Tables III-VIII), not by a formal argument that makes the outcome true by definition. No target quantity is defined in terms of itself: the classifier outputs are patch labels, and the reported accuracy, sensitivity, specificity, F1-score, and AUC are computed from confusion matrices. The literature citations used to motivate selective fine-tuning ([25], [30], [32]) are external studies, not self-citations, and none is invoked to forbid alternative architectures. The KMC dataset is attributed to Aatresh et al. [2], but that is data reuse, not a self-citation chain. The only circularity-adjacent issue is that the best feature extractor is selected after evaluating all eight architectures on the same test split that is later used for reporting (Tables III-VI), and the patch-level split does not group by whole-slide image; these are statistical validity risks that can inflate the reported numbers, but they do not make the 'hybrid helps' conclusion true by construction. The stated limitations discuss parameter count only and do not mention these hazards. Under the requirement to flag only explicit reductions to inputs, no circular step is present.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The paper contributes no new mathematical derivation and no new entities. Its claims rest on standard transfer-learning assumptions and several empirical choices, the most fragile being the independence of patch-level splits and the use of the test set for model selection. The free parameters are mostly standard training hyperparameters plus the empirically chosen deep classifier depth.

free parameters (5)
  • Deep classifier architecture (number and width of fully connected layers) = Not stated (decided empirically)
    Section II.B says 'The number of additional fully connected layers is decided empirically.'
  • Number of fine-tuned top layers in the feature extractor = Not stated
    The paper says 'selective top layers' are trainable but never specifies how many or which layers.
  • Patch selection thresholds = Mean intensity 200, standard deviation 60
    Used to filter valid patches during extraction (Section II.A); chosen empirically.
  • Best feature extractor per dataset = ResNet50 for TCGA, EfficientNetb3 for KMC, EfficientNetb2 for COLON
    Selected as the maximum over eight models evaluated on the respective test sets, a post-hoc model selection.
  • Training hyperparameters = lr=0.001, batch size 64/32, epochs 47, warm restart 12, Adam
    Reported in Table II; no sensitivity analysis is performed.
assumptions (5)
  • domain assumption ImageNet pre-trained features, after fine-tuning of selected top layers, transfer usefully to H&E histopathology images.
    The entire transfer-learning strategy rests on this; the paper provides no comparison with training from scratch.
  • domain assumption Random patch-level splitting of whole slide images yields statistically independent training and test sets.
    Table I and Section II.C describe splitting at the patch level; patches from the same slide are likely correlated, so this assumption is fragile and load-bearing.
  • domain assumption Patch labels inherit the slide-level class labels correctly; every patch within a WSI is a valid representative of that class.
    No pathologist re-annotation of individual patches is described; mixed-content patches could be mislabeled.
  • domain assumption Macenko color normalization adequately removes staining variability across datasets.
    Applied to TCGA patches; KMC patches were received already pre-processed. The normalization's benefit is assumed, not ablated.
  • domain assumption Weighted random sampling corrects class imbalance without biasing the learned features.
    Used to balance classes; the paper does not analyze its effect on validation performance.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hybrid deep learning-based strategy for the hepatocellular carcinoma cancer grade classification of H&E stained liver histopathology images." pith.science (2026). https://pith.science/paper/DDN2MVEZ

@misc{pith2026241203084,
  author       = {Pith},
  title        = {Pith review of: Hybrid deep learning-based strategy for the hepatocellular carcinoma cancer grade classification of H&E stained liver histopathology images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DDN2MVEZ}},
  note         = {Machine review of arXiv:2412.03084}
}
read the original abstract

Hepatocellular carcinoma (HCC) is a common type of liver cancer whose early-stage diagnosis is a common challenge, mainly due to the manual assessment of hematoxylin and eosin-stained whole slide images, which is a time-consuming process and may lead to variability in decision-making. For accurate detection of HCC, we propose a hybrid deep learning-based architecture that uses transfer learning to extract the features from pre-trained convolutional neural network (CNN) models and a classifier made up of a sequence of fully connected layers. This study uses a publicly available The Cancer Genome Atlas Hepatocellular Carcinoma (TCGA-LIHC)database (n=491) for model development and database of Kasturba Gandhi Medical College (KMC), India for validation. The pre-processing step involves patch extraction, colour normalization, and augmentation that results in 3920 patches for the TCGA dataset. The developed hybrid deep neural network consisting of a CNN-based pre-trained feature extractor and a customized artificial neural network-based classifier is trained using five-fold cross-validation. For this study, eight different state-of-the-art models are trained and tested as feature extractors for the proposed hybrid model. The proposed hybrid model with ResNet50-based feature extractor provided the sensitivity, specificity, F1-score, accuracy, and AUC of 100.00%, 100.00%, 100.00%, 100.00%, and 1.00, respectively on the TCGA database. On the KMC database, EfficientNetb3 resulted in the optimal choice of the feature extractor giving sensitivity, specificity, F1-score, accuracy, and AUC of 96.97, 98.85, 96.71, 96.71, and 0.99, respectively. The proposed hybrid models showed improvement in accuracy of 2% and 4% over the pre-trained models in TCGA-LIHC and KMC databases.

Figures

Figures reproduced from arXiv: 2412.03084 by the authors.

Figure 1
Figure 1. Block diagram showing overall workflow of the experimentation [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. TCGA-LIHC liver HCC database [1]. This diagram describes the [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. TCGA-LIHC, KMC and COLON datasets are comprised of 3, 4 [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (10 more)
Figure 2
Figure 2. Figure 2: This study also uses another cancer database of liver HCC collected from Kasturba Gandhi Medical College, India [2]. The KMC database has four types of liver HCC tumors such as type 0, type 1, type 2, and type 3 having 719, 799, 776, and 711 image patches, respectively…
Figure 5
Figure 5. Figure 5: The workflow of the training and validation of the proposed method [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: Train and validation loss plotted for all epochs, all 5 folds over [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: Train and validation accuracy plotted for all epochs, all 5 folds [PITH_FULL_IMAGE:figures/full_fig_p005_7.png]
Figure 8
Figure 8. Figure 8: Confusion matrix of all 5 folds over TCGA test dataset of trained [PITH_FULL_IMAGE:figures/full_fig_p005_8.png]
Figure 11
Figure 11. Figure 11: Training and validation accuracy plotted for all epochs, all 5 [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]
Figure 12
Figure 12. Figure 12: Confusion matrix of all 5 folds over KMC test dataset of trained [PITH_FULL_IMAGE:figures/full_fig_p008_12.png]
Figure 13
Figure 13. Figure 13: ROC curve of all 5 folds over KMC test dataset of trained hybrid [PITH_FULL_IMAGE:figures/full_fig_p008_13.png]
Figure 14
Figure 14. Figure 14: Confusion matrix of all 5 folds over COLON test dataset of trained [PITH_FULL_IMAGE:figures/full_fig_p009_14.png]
Figure 15
Figure 15. Figure 15: ROC curve of all 5 folds over COLON test dataset of trained [PITH_FULL_IMAGE:figures/full_fig_p009_15.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 37 canonical work pages

  1. [1]

    The cancer genome atlas liver hepatocellular carcinoma collection (TCGA-LIHC),

    B. J. Erickson, S. Kirk, Y . Lee, O. Bathe, M. Kearns, C. Gerdes, K. Rieger-Christ, and J. Lemmerman, “The cancer genome atlas liver hepatocellular carcinoma collection (TCGA-LIHC),” 2016. 11

    work page 2016

  2. [2]

    Livernet: efficient and robust deep learning model for automatic diagnosis of sub- types of liver hepatocellular carcinoma cancer from h&e stained liver histopathology images,

    A. A. Aatresh, K. Alabhya, S. Lal, J. Kini, and P. Saxena, “Livernet: efficient and robust deep learning model for automatic diagnosis of sub- types of liver hepatocellular carcinoma cancer from h&e stained liver histopathology images,” Int J Comput Assist Radiol Surg , vol. 16, no. 9, pp. 1549–1563, 2021

    work page 2021

  3. [3]

    Hepatocellular carcinoma: epi- demiology and molecular carcinogenesis,

    H. B. El-Serag and K. L. Rudolph, “Hepatocellular carcinoma: epi- demiology and molecular carcinogenesis,” Gastroenterology, vol. 132, pp. 2557–2576, June 2007

    work page 2007

  4. [4]

    Key statistics about liver cancer

    “Key statistics about liver cancer.”

    work page

  5. [5]

    Global, regional and national burden of primary liver cancer by subtype,

    H. Rumgay, J. Ferlay, C. de Martel, D. Georges, A. S. Ibrahim, R. Zheng, W. Wei, V . E. Lemmens, and I. Soerjomataram, “Global, regional and national burden of primary liver cancer by subtype,” Eur . J. Cancer ., vol. 161, pp. 108–118, 2022

    work page 2022

  6. [6]

    Staged detection–identification framework for cell nuclei in histopathology images,

    X. Li, W. Li, and R. Tao, “Staged detection–identification framework for cell nuclei in histopathology images,” IEEE Transactions on Instru- mentation and Measurement , vol. 69, no. 1, pp. 183–193, 2019

    work page 2019

  7. [7]

    Classification of multi-differentiated liver cancer pathological images based on deep learning attention mechanism,

    C. Chen, C. Chen, M. Ma, X. Ma, X. Lv, X. Dong, Z. Yan, M. Zhu, and J. Chen, “Classification of multi-differentiated liver cancer pathological images based on deep learning attention mechanism,” BMC Med Inform Decis Mak , vol. 22, p. 176, July 2022

    work page 2022

  8. [8]

    Clinical impact and frequency of anatomic pathology errors in cancer diagnoses,

    S. S. Raab, D. M. Grzybicki, J. E. Janosky, R. J. Zarbo, F. A. Meier, C. Jensen, and S. J. Geyer, “Clinical impact and frequency of anatomic pathology errors in cancer diagnoses,” Cancer, vol. 104, pp. 2205–2213, Nov. 2005

    work page 2005

Show all 43 references
  1. [9]

    Diagnostic concordance among pathologists interpreting breast biopsy specimens,

    J. G. Elmore, G. M. Longton, P. A. Carney, B. M. Geller, T. Onega, A. N. A. Tosteson, H. D. Nelson, M. S. Pepe, K. H. Allison, S. J. Schnitt, F. P. O’Malley, and D. L. Weaver, “Diagnostic concordance among pathologists interpreting breast biopsy specimens,” JAMA, vol. 313, pp....

    2015

  2. [10]

    Error reduction in surgical pathology,

    R. E. Nakhleh, “Error reduction in surgical pathology,” Arch Pathol Lab Med, vol. 130, pp. 630–632, May 2006

    2006

  3. [11]

    Guest editorial deep learning in medical imaging: Overview and future promise of an exciting new technique,

    H. Greenspan et al., “Guest editorial deep learning in medical imaging: Overview and future promise of an exciting new technique,” IEEE Transactions on Medical Imaging , vol. 35, no. 5, pp. 1153–1159, 2016

    2016

  4. [12]

    Deep learning- based classification of liver cancer histopathology images using only global labels,

    C. Sun, A. Xu, D. Liu, Z. Xiong, F. Zhao, and W. Ding, “Deep learning- based classification of liver cancer histopathology images using only global labels,” IEEE J. Biomed. Health Inform. , vol. 24, no. 6, pp. 1643– 1651, 2019

    2019

  5. [13]

    Classification and mutation prediction based on histopathology h&e images in liver cancer using deep learning,

    M. Chen, B. Zhang, W. Topatana, J. Cao, H. Zhu, S. Juengpanich, Q. Mao, H. Yu, and X. Cai, “Classification and mutation prediction based on histopathology h&e images in liver cancer using deep learning,” npj Precision Oncology, vol. 4, p. 14, Jun 2020

    2020

  6. [14]

    Support-vector networks,

    C. Cortes and V . Vapnik, “Support-vector networks,” Machine learning, vol. 20, no. 3, pp. 273–297, 1995

    1995

  7. [15]

    Imagenet classification with deep convolutional neural networks,

    A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,” vol. 60, pp. 84–90, AcM New York, NY , USA, 2017

    2017

  8. [16]

    Very deep convolutional networks for large-scale image recognition,

    K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” 2015

    2015

  9. [17]

    Deep residual learning for image recognition,

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

    2016

  10. [18]

    Densely connected convolutional networks,

    G. Huang, Z. Liu, L. van der Maaten, and K. Q. Weinberger, “Densely connected convolutional networks,” in Proceedings of the IEEE Confer- ence on Computer Vision and Pattern Recognition (CVPR) , pp. 4700– 4708, 2017

    2017

  11. [19]

    EfficientNet: Rethinking model scaling for con- volutional neural networks,

    M. Tan and Q. Le, “EfficientNet: Rethinking model scaling for con- volutional neural networks,” in Proceedings of the 36th International Conference on Machine Learning , pp. 6105–6114, 2019

    2019

  12. [20]

    BreastNet: A novel convolutional neural network model through histopathological images for the diagnosis of breast cancer,

    M. To ˘gaçar, K. B. Özkurt, B. Ergen, and Z. Cömert, “BreastNet: A novel convolutional neural network model through histopathological images for the diagnosis of breast cancer,” Physica A , vol. 545, p. 123592, 2020

    2020

  13. [21]

    A deep learning method for breast cancer classification in the pathology images,

    M. Liu et al. , “A deep learning method for breast cancer classification in the pathology images,” IEEE Journal of Biomedical and Health Informatics, vol. 26, no. 10, pp. 5025–5032, 2022

    2022

  14. [22]

    Deep learning vs. traditional computer vision,

    N. O’Mahony, S. Campbell, A. Carvalho, S. Harapanahalli, G. V . Hernandez, L. Krpalkova, D. Riordan, and J. Walsh, “Deep learning vs. traditional computer vision,” pp. 128–144, 2020

    2020

  15. [23]

    Grading of hcc biopsy images using nucleus and texture features,

    G. Chakraborty et al., “Grading of hcc biopsy images using nucleus and texture features,” IEEE Journal of Biomedical and Health Informatics , vol. 27, no. 1, pp. 65–74, 2023

    2023

  16. [24]

    Dca-daffnet: An end-to-end network with deformable fusion attention and deep adaptive feature fusion for laryn- geal tumor grading from histopathology images,

    J. Luo, P. Huang, P. He, B. Wei, X. Guo, H. Xiao, Y . Sun, S. Tian, M. Zhou, and P. Feng, “Dca-daffnet: An end-to-end network with deformable fusion attention and deep adaptive feature fusion for laryn- geal tumor grading from histopathology images,” IEEE Transactions on Instr...

    2023

  17. [25]

    How deeply to fine-tune a convolutional neural network: A case study using a histopathology dataset,

    I. Kandel and M. Castelli, “How deeply to fine-tune a convolutional neural network: A case study using a histopathology dataset,” Appl. Sci., vol. 10, no. 10, pp. 1–20, 2020

    2020

  18. [26]

    Breast cancer histology images classification: Training from scratch or transfer learning?,

    Shallu and R. Mehra, “Breast cancer histology images classification: Training from scratch or transfer learning?,” ICT Express, vol. 4, no. 4, pp. 247–254, 2018

    2018

  19. [27]

    Unsupervised feature extraction via deep learning for histopathological classification of colon tissue images,

    C. T. Sari et al. , “Unsupervised feature extraction via deep learning for histopathological classification of colon tissue images,” IEEE Transac- tions on Medical Imaging , vol. 38, no. 5, pp. 1139–1149, 2018

    2018

  20. [28]

    A study on cnn transfer learning for image classification,

    M. Hussain, J. J. Bird, and D. R. Faria, “A study on cnn transfer learning for image classification,” in UK Workshop on Computational Intelligence, pp. 191–202, Springer, 2018

    2018

  21. [29]

    Automated classification of histopathology images using transfer learning,

    M. Talo, “Automated classification of histopathology images using transfer learning,” Artif. Intell. Med. , vol. 101, no. 101743, pp. 1–10, 2019

    2019

  22. [30]

    Application of deep transfer learning for automated brain abnormality classification using MR images,

    M. Talo, U. B. Baloglu, Ö. Yıldırım, and U. Rajendra Acharya, “Application of deep transfer learning for automated brain abnormality classification using MR images,” Cogn. Syst. Res. , vol. 54, pp. 176–188, May 2019

    2019

  23. [31]

    Lesion classification of coronary artery cta images based on cbam and transfer learning,

    Y . Jin, X. Ye, N. Feng, Z. Wang, X. Hei, J. Liu, L. Mu, and Y . Li, “Lesion classification of coronary artery cta images based on cbam and transfer learning,” IEEE Transactions on Instrumentation and Measurement , vol. 73, pp. 1–14, 2024

    2024

  24. [32]

    How transferable are features in deep neural networks?,

    J. Yosinski, J. Clune, Y . Bengio, and H. Lipson, “How transferable are features in deep neural networks?,” Advances in neural information processing systems, vol. 27, 2014

    2014

  25. [33]

    Multi-phase fine-tuning: A new fine-tuning approach for sign language recognition,

    N. Sarhan, M. Lauri, and S. Frintrop, “Multi-phase fine-tuning: A new fine-tuning approach for sign language recognition,” KI - Künstliche Intelligenz, vol. 36, pp. 91–98, Mar 2022

    2022

  26. [34]

    Convolutional neural networks for medical image analysis: Full training or fine tuning?,

    N. Tajbakhsh et al. , “Convolutional neural networks for medical image analysis: Full training or fine tuning?,” IEEE Transactions on Medical Imaging, vol. 35, no. 5, pp. 1299–1312, 2016

    2016

  27. [35]

    A method for normalizing histology slides for quantitative analysis,

    M. Macenko, M. Niethammer, J. S. Marron, D. Borland, J. T. Woosley, X. Guan, C. Schmitt, and N. E. Thomas, “A method for normalizing histology slides for quantitative analysis,” in IEEE International Sym- posium on Biomedical Imaging: From Nano to Macro , pp. 1107–1110, 2009

    2009

  28. [36]

    Lung and colon cancer histopathological image dataset (lc25000),

    A. A. Borkowski, M. M. Bui, L. B. Thomas, C. P. Wilson, L. A. DeLand, and S. M. Mastorides, “Lung and colon cancer histopathological image dataset (lc25000),” 2019

    2019

  29. [37]

    Pytorch: An imperative style, high- performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, “Pytorch: An imperative style, high- p...

    2019

  30. [38]

    Sgdr: Stochastic gradient descent with warm restarts,

    I. Loshchilov and F. Hutter, “Sgdr: Stochastic gradient descent with warm restarts,” 2017

    2017

  31. [39]

    A weakly supervised method with colorization for nuclei segmentation using point annotations,

    L. Xia, Z. Qu, J. An, and Z. Gao, “A weakly supervised method with colorization for nuclei segmentation using point annotations,” IEEE Transactions on Instrumentation and Measurement , vol. 72, pp. 1–11, 2023

    2023

  32. [40]

    Metrolog- ical characterization of a cadx system for the classification of breast masses in mammograms,

    A. Mencattini, M. Salmeri, G. Rabottino, and S. Salicone, “Metrolog- ical characterization of a cadx system for the classification of breast masses in mammograms,” IEEE Transactions on Instrumentation and Measurement, vol. 59, no. 11, pp. 2792–2799, 2010

    2010

  33. [41]

    A dataset for breast cancer histopathological image classification,

    F. A. Spanhol, L. S. Oliveira, C. Petitjean, and L. Heutte, “A dataset for breast cancer histopathological image classification,” IEEE Transactions on Biomedical Engineering , vol. 63, no. 7, pp. 1455–1462, 2016

    2016

  34. [42]

    Deep feature selection using adaptive β-hill climbing aided whale optimization algorithm for lung and colon cancer detection,

    A. Bhattacharya, B. Saha, S. Chattopadhyay, and R. Sarkar, “Deep feature selection using adaptive β-hill climbing aided whale optimization algorithm for lung and colon cancer detection,” Biomedical Signal Processing and Control , vol. 83, p. 104692, 2023

    2023

  35. [43]

    Machine learning-based lung and colon cancer detection using deep feature extraction and ensemble learning,

    M. A. Talukder, M. M. Islam, M. A. Uddin, A. Akhter, K. F. Hasan, and M. A. Moni, “Machine learning-based lung and colon cancer detection using deep feature extraction and ensemble learning,” Expert Systems with Applications , vol. 205, p. 117695, 2022

    2022

Pith tools

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