Enhancing Transfer Learning for Medical Image Classification with SMOTE: A Comparative Study

H.M. Nahid Kawsar; Iffat Rimi; Md. Zehan Alam; Tonmoy Roy

REVIEW 3 major objections 4 minor 16 references

Enhancing Transfer Learning for Medical Image Classification with SMOTE: A Comparative Study

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read SMOTE-plus-transfer-learning lifts diabetic retinopathy recall by 5.4 points.

desk verdict A useful but confounded empirical result: the reported SMOTE gains on APTOS need an ablation before causal credit. read the letter →

arxiv 2412.20235 v1 pith:UY5OA3YJ submitted 2024-12-28 eess.IV cs.CV

Md. Zehan Alam , Tonmoy Roy , H.M. Nahid Kawsar , Iffat Rimi This is my paper · ORCID

classification eess.IVcs.CV

keywords transferlearningSMOTEdiabeticretinopathybraintumorMRIclassimbalanceensemblemedicalimageclassificationXception

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 tests whether transfer-learned CNN features, rebalanced with SMOTE and fed to classical machine-learning classifiers, can fix the sensitivity shortfall of transfer learning alone on imbalanced medical images. On the APTOS 2019 diabetic retinopathy dataset it reports that the best hybrid, a Voting Classifier trained on SMOTE-resampled Xception probabilities, reaches 92.14% accuracy and 85.50% recall versus 90.17% and 80.06% for Xception alone. On the balanced brain tumor MRI dataset, by contrast, transfer learning is already near-perfect and SMOTE adds nothing. The practical stake is a computationally light way to improve detection of underrepresented disease classes without retraining large networks.

What carries the argument

The carried mechanism is a three-stage pipeline: a frozen ImageNet pre-trained CNN (Xception for the diabetic retinopathy set) outputs class probabilities for each training image; SMOTE synthesizes new probability-vector samples for minority classes by interpolating between existing ones; then classical classifiers, configured with balanced class weights, are trained on the resampled probability vectors. This converts an imbalanced image-classification problem into a balanced tabular classification problem, avoiding the cost of oversampling raw images.

What would settle it

Holding Xception probabilities and the Voting Classifier fixed, retrain once with SMOTE and once without it, keeping balanced class weights in both runs; if recall stays near 85% without SMOTE, the reported gain is not attributable to SMOTE.

Watch

Extended reading notes

Core claim

On imbalanced medical image data, replacing the transfer-learned CNN's final classifier with a classical machine-learning pipeline trained on SMOTE-resampled CNN probabilities improves sensitivity more than accuracy. The paper's central result is on the five-class APTOS 2019 diabetic retinopathy set: Xception transfer learning alone gives 90.17% accuracy and 80.06% recall, while the Voting Classifier (Logistic Regression, Linear SVM, SGD, Extra Trees, Random Forest) on SMOTE-balanced Xception probabilities gives 92.14% accuracy and 85.50% recall, with specificity rising from 97.46% to 98.18%. The same pipeline leaves the balanced brain tumor benchmark essentially unchanged, which the paper takes as evidence that resampling matters only when class imbalance is the bottleneck.

Load-bearing premise

The argument that SMOTE produces the improvement assumes nothing else in the final pipeline changed, but the comparison swaps the classifier for an ensemble and turns on balanced class weights at the same time.

Editorial extensions

If this is right

If this holds, transfer learning alone is sufficient on near-balanced medical datasets, and SMOTE-based rebalancing is only needed when class imbalance dominates.
The reported 5.44-point recall gain on the APTOS dataset means more true cases of diabetic retinopathy would be caught than with the transfer-learning baseline, at minimal added compute.
The hybrid approach can be applied to any pre-trained CNN by saving its training-set probabilities, so it does not require re-training the deep network.
The small accuracy gain (1.97 points) alongside the larger recall gain suggests the method trades little overall correctness for much better sensitivity.
Because the BT results show no degradation after SMOTE, the method appears safe on balanced data as well, though it offers no gain there.

Reading between the lines

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

The paper's attribution of the gain to SMOTE is underdetermined: the final comparison changes three variables at once (the classifier type, the balanced class-weight setting, and SMOTE), so an ablation holding the classifier and class weights fixed would be needed to isolate SMOTE's true contribution.
SMOTE here is applied to CNN probability vectors, not to raw images; a testable extension is whether oversampling raw image embeddings rather than probabilities would yield larger gains or better calibration.
The claim that no test image was falsely labeled 'No DR' when it actually had disease is class-specific; a per-class confusion matrix would show whether the sensitivity gain is concentrated in particular severity stages.
The same probability-plus-SMOTE recipe could be tried on other imbalanced medical datasets (e.g., rare cancers or lesion types) where a pre-trained CNN is already available but its raw sensitivity is too low for screening use.

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.

Desk Editor's Note

read the letter

Colleague,

The paper is a straightforward empirical comparison: five ImageNet CNNs fine-tuned on two public medical datasets, then class probabilities fed to classical ML classifiers with SMOTE. The one concrete new number is on APTOS: the Voting Classifier on SMOTE-resampled Xception probabilities gets 92.14% accuracy and 85.50% recall versus 90.17% and 80.06% for Xception fine-tuning alone. That is a plausible and potentially useful gain for screening, and it costs little compute.

Credit where due: the experiments are reproducible in shape (public datasets, standard models), the metric tables are internally consistent, and the authors explicitly note that SMOTE did nothing on the balanced brain-tumor dataset, which is the kind of negative result that lends the work some credibility.

However, the central attribution is not isolated. The hybrid pipeline changes three things at once: the classifier becomes a five-model voting ensemble, 'balanced' class weights are applied where available, and SMOTE is applied. The abstract says SMOTE improves recall by 5.43%, but the comparison in Tables V and VI is Xception-only versus Voting+SMOTE. An ablation (Xception probabilities + Voting without SMOTE, Xception probabilities + SMOTE + single ML, etc.) is needed before crediting SMOTE. Given that SMOTE gave nothing on the BT dataset, it may well be the ensemble or class weighting doing the work.

There is also a leakage-adjacent design issue: the ML stage trains on in-sample CNN probabilities with no out-of-fold scheme. SMOTE interpolates in a probability simplex geometry that may not reflect test-time predictions. No error bars or significance tests either, so the 5-point recall gain could be within run-to-run noise.

All that said, this is a working-paper-level contribution that a serious referee could improve with a 2x2 ablation and a nested cross-validation. It is not a conceptual breakthrough, but it is an honest measurement worth reporting. I would send it to peer review expecting revision. I would not cite it for the SMOTE claim until the ablation is done.

Referee Report

3 major / 4 minor

Summary. The paper evaluates five ImageNet-pretrained CNN models (MobileNet, Xception, InceptionV3, ResNet50, DenseNet201) on two medical imaging datasets: a brain tumor MRI dataset and the APTOS 2019 diabetic retinopathy dataset. For the brain tumor dataset, transfer learning alone achieves near-perfect results. For the imbalanced diabetic retinopathy dataset, the authors propose a hybrid pipeline in which CNN class probabilities are used as features, SMOTE is applied to balance the training set, and traditional machine learning classifiers are trained on the resampled features. They report that a Voting Classifier trained on SMOTE-resampled Xception probabilities improves accuracy from 90.17% to 92.14%, recall from 80.06% to 85.50%, and specificity from 97.46% to 98.18%, compared with Xception transfer learning alone. The paper concludes that combining transfer learning with SMOTE and traditional classifiers is a computationally light way to mitigate class imbalance in medical image classification.

Significance. If the attribution to SMOTE were established, the paper would offer a practically useful and computationally inexpensive post-processing method for improving sensitivity on imbalanced medical imaging datasets. The study has clear strengths: it uses two public datasets, compares several CNN architectures, and reports a consistent set of classification metrics. The internal consistency of Tables III–VI is good, and the observation that SMOTE provides no benefit on the already balanced brain tumor dataset is a useful negative control. However, the central claim is not yet causally supported because the improvement is measured between a single CNN baseline and a pipeline that changes multiple components at once, with no ablation, no repeated runs, and no statistical uncertainty assessment. The core result is plausible but currently under-identified.

major comments (3)

[IV.B, Tables V and VI; Abstract] The reported 1.97% accuracy gain and 5.44% recall gain are attributed to SMOTE, but the comparison changes three variables simultaneously: the single Xception classifier is replaced by a five-model Voting Classifier, class weights are set to 'balanced' where available (Section III.F), and SMOTE is applied (Section III.E). A 2x2 ablation is needed—SMOTE on/off crossed with class weighting on/off, and ideally with a single logistic regression as well as the ensemble—to determine whether the recall improvement comes from SMOTE, from the balanced class weights, or from ensemble averaging. Without this ablation, the abstract's causal statement that 'we integrate SMOTE... which improves accuracy by 1.97%' is not justified by the experimental design.
[III.D–III.E] The machine-learning stage is trained on in-sample CNN probabilities: Section III.D states that 'predicted class probabilities for training and test datasets were retained,' and Section III.E says these training probabilities form the new dataset. No out-of-fold or cross-validated probability generation is described. Because SMOTE interpolates between training points in probability space, synthetic samples may lie in regions that do not correspond to test-time CNN outputs, especially given that the same training set was used to fit the CNN. The authors should generate the feature set using out-of-fold predictions (or at least evaluate with an inner cross-validation loop) to avoid optimistic bias in the SMOTE + ML stage.
[IV (overall evaluation protocol)] All reported numbers come from a single training run, with no error bars, confidence intervals, or significance tests. The improvement from 90.17% to 92.14% may be within stochastic variation, particularly because training uses stochastic augmentation and early stopping. In addition, Section IV says 'We'll only use the top 5 of the final results to evaluate,' which implies that the best of many ML models was selected on the test set; this selection effect is not accounted for in the reported gains. The authors should report repeated runs or use a nested validation scheme so that the headline improvement is not an artifact of model selection or random seed.

minor comments (4)

[Abstract and throughout] The paper uses 'multilabel image classification' but the datasets are multiclass, single-label problems; this terminology should be corrected.
[III.B] The text says the ImageDataGenerator class comes from scikit-learn, but ImageDataGenerator is provided by Keras/TensorFlow, not scikit-learn.
[III.A] The narrative switches to first person ('I further divided the remaining 20%') in a multi-author paper; this should be 'we' throughout.
[V, Discussion] The sentence 'the moderate class was predominantly misclassified, particularly as Proliferate and Severe classes' appears to contradict Figure 5, which likely shows misclassifications into adjacent stages. Please verify the confusion matrix interpretation and clarify that misclassification into more severe stages is not necessarily clinically 'less critical.'

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; the SMOTE improvement claim is an empirical comparison with a confounded baseline, not a derivation.

full rationale

The paper's central claim is an empirical performance comparison, not a mathematical derivation. No parameter is fitted and then renamed as a prediction, no quantity is defined in terms of the claimed result, and no load-bearing conclusion rests on a self-citation: all cited works are external, and the authors do not cite their own prior results as a basis for the method. The pipeline does change several variables at once—the single Xception transfer-learning classifier is replaced by a Voting Classifier over multiple ML models, SMOTE is applied, and 'balanced' class weights are enabled in Section III.F—so the reported accuracy and recall gains cannot be causally attributed to SMOTE alone. That is a genuine experimental confound and an omission of an ablation study, but it is not circular reasoning under the definitions used here: the numbers are direct measurements, and the paper does not assert that SMOTE's effect follows by construction from its own equations. The in-sample use of CNN training probabilities as features for the ML stage is standard stacking practice and creates a risk of optimistic evaluation, but it is not proof-by-assumption. Therefore no circular step is present, and the appropriate score is 0.

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

The paper introduces no new mathematical objects. It relies on established algorithms (CNN architectures, SMOTE, classical classifiers) and several unstated assumptions about feature sufficiency and data representativeness. The main unquantified choices are the ML model hyperparameters and SMOTE settings, which are not reported.

free parameters (3)

ML classifier hyperparameters (e.g., Voting Classifier, SGD, SVM, Naive Bayes) = not reported
The paper lists which algorithms are used but does not report the hyperparameters or whether they were tuned. These choices affect the reported metrics and are not specified.
SMOTE sampling strategy and k_neighbors = not reported
SMOTE requires choosing the amount of oversampling (e.g., whether all classes are balanced to the majority count) and the number of nearest neighbors. These settings determine the synthetic training set and are not disclosed.
Data augmentation parameters (rotation up to 60 degrees, brightness 0.8-1.2) = chosen by hand
The augmentation ranges are chosen by hand and affect the CNN probabilities used as features, but they are standard choices and secondary to the central claim.

assumptions (4)

domain assumption ImageNet-pretrained weights transfer effectively to brain MRI and retinal fundus images.
The entire TL pipeline assumes the features learned on natural images are useful for medical images; the paper cites prior work but does not validate this independently.
domain assumption The softmax class probabilities from the CNN form a sufficient feature representation for downstream classical classifiers.
The paper uses CNN training probabilities directly as features for ML models without exploring other representations or validating this choice (Section III.E).
domain assumption SMOTE-generated synthetic samples follow the same distribution as real minority-class samples.
SMOTE interpolates between existing minority samples; the paper assumes these synthetic examples improve classifier training without being misleading (Section III.E).
domain assumption A single random split into train/validation/test is representative.
No cross-validation or repeated splits are used; the reported numbers depend on one split of the APTOS dataset.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Enhancing Transfer Learning for Medical Image Classification with SMOTE: A Comparative Study." pith.science (2026). https://pith.science/paper/UY5OA3YJ

@misc{pith2026241220235,
  author       = {Pith},
  title        = {Pith review of: Enhancing Transfer Learning for Medical Image Classification with SMOTE: A Comparative Study},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UY5OA3YJ}},
  note         = {Machine review of arXiv:2412.20235}
}

read the original abstract

This paper explores and enhances the application of Transfer Learning (TL) for multilabel image classification in medical imaging, focusing on brain tumor class and diabetic retinopathy stage detection. The effectiveness of TL-using pre-trained models on the ImageNet dataset-varies due to domain-specific challenges. We evaluate five pre-trained models-MobileNet, Xception, InceptionV3, ResNet50, and DenseNet201-on two datasets: Brain Tumor MRI and APTOS 2019. Our results show that TL models excel in brain tumor classification, achieving near-optimal metrics. However, performance in diabetic retinopathy detection is hindered by class imbalance. To mitigate this, we integrate the Synthetic Minority Over-sampling Technique (SMOTE) with TL and traditional machine learning(ML) methods, which improves accuracy by 1.97%, recall (sensitivity) by 5.43%, and specificity by 0.72%. These findings underscore the need for combining TL with resampling techniques and ML methods to address data imbalance and enhance classification performance, offering a pathway to more accurate and reliable medical image analysis and improved patient outcomes with minimal extra computation powers.

Figures

Figures reproduced from arXiv: 2412.20235 by the authors.

**Figure 2.** Sample Images from Each Class of DR Dataset [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Workflow of the Proposed Model for Medical Image Classification [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Confusion Matrix of Final Prediction on BT Dataset [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

16 extracted references · 13 canonical work pages

[1]

S. G. De Benedictis, G. Gargano, and G. Settembre, ”Enhanced MRI brain tumor detection and classification via topological data analy- sis and low-rank tensor decomposition,” Journal of Computational Mathematics and Data Science , vol. 13, 2024, Art. no. 100103. doi: 10.1016/j.jcmds.2024.100103

arXiv 2024
[2]

Sikder, M

N. Sikder, M. S. Chowdhury, A. S. M. Arif, and A.-A. Nahid, ”Early blindness detection based on retinal images using ensemble learning,” in Proc. 2019 22nd Int. Conf. Comput. Inf. Technol. (ICCIT) , Dhaka, Bangladesh, 2019, pp. 1-6. doi: 10.1109/ICCIT48885.2019.9038439

work page arXiv 2019
[3]

Breast Cancer Detection and Localizing the Mass Area Using Deep Learning,

M. M. Rahman et al., “Breast Cancer Detection and Localizing the Mass Area Using Deep Learning,” Big Data Cogn. Comput. , vol. 8, no. 80, 2024, doi: 10.3390/bdcc8070080

work page doi:10.3390/bdcc8070080 2024
[4]

Two-Stage Selective Ensemble of CNN via Deep Tree Training for Medical Image Classification,

Y . Yang, Y . Hu, X. Zhang, and S. Wang, “Two-Stage Selective Ensemble of CNN via Deep Tree Training for Medical Image Classification,” IEEE Trans. Cybern. , vol. 52, no. 9, pp. 9194–9207, Sep. 2022, doi: 10.1109/TCYB.2021.3061147

arXiv 2022
[5]

Feature fusion and Ensemble learning-based CNN model for mammographic image classification,

I. Ul Haq, H. Ali, H. Y . Wang, C. Lei, and H. Ali, “Feature fusion and Ensemble learning-based CNN model for mammographic image classification,” J. King Saud Univ. Comput. Inf. Sci. , vol. 34, no. 6, Part B, pp. 3310–3318, 2022, doi: 10.1016/j.jksuci.2022.03.023

work page doi:10.1016/j.jksuci.2022.03.023 2022
[6]

Lung can- cer medical images classification using hybrid CNN-SVM,

A. Y . Saleh, C. K. Chin, V . Penshie, and H. Al-Absi, “Lung can- cer medical images classification using hybrid CNN-SVM,” Int. J. Adv. Intell. Informatics , vol. 7, no. 2, pp. 151–162, Jul. 2021, doi: 10.26555/ijain.v7i2.317

work page doi:10.26555/ijain.v7i2.317 2021
[7]

Transfer Learning in Breast Cancer Diagnoses via Ultrasound Imaging,

G. Ayana, K. Dese, and S.-w. Choe, “Transfer Learning in Breast Cancer Diagnoses via Ultrasound Imaging,” Cancers, vol. 13, no. 738, 2021, doi: 10.3390/cancers13040738

work page doi:10.3390/cancers13040738 2021
[8]

Towards a Better Understanding of Transfer Learning for Medical Imaging: A Case Study,

L. Alzubaidi et al., “Towards a Better Understanding of Transfer Learning for Medical Imaging: A Case Study,” Applied Sciences , vol. 10, no. 13, pp. 4523, 2020, doi: 10.3390/app10134523

work page doi:10.3390/app10134523 2020

Show all 16 references

[9]

An Analysis on Ensemble Learn- ing optimized Medical Image Classification with Deep Convolutional Neural Networks,

D. M ¨uller, I. Soto-Rey, and F. Kramer, “An Analysis on Ensemble Learn- ing optimized Medical Image Classification with Deep Convolutional Neural Networks,” 2022

2022
[10]

Neu- roNet19: an explainable deep neural network model for the classification of brain tumors using magnetic resonance imaging data,

R. Haque, M. M. Hassan, A. K. Bairagi, and S. M. Shariful Islam, “Neu- roNet19: an explainable deep neural network model for the classification of brain tumors using magnetic resonance imaging data,” Sci. Rep., vol. 14, no. 1, pp. 1524, Jan. 2024, doi: 10.1038/s41598-024-51867-1

2024 doi
[11]

MRI-Based Brain Tumor Classification Using Ensemble of Deep Features and Machine Learning Classifiers,

J. Kang, Z. Ullah, and J. Gwak, “MRI-Based Brain Tumor Classification Using Ensemble of Deep Features and Machine Learning Classifiers,” Sensors, vol. 21, no. 6, pp. 2222, Mar. 2021, doi: 10.3390/s21062222

2021 doi
[12]

An Effective Approach to Detect and Identify Brain Tumors Using Transfer Learning,

N. Ullah et al., “An Effective Approach to Detect and Identify Brain Tumors Using Transfer Learning,” Applied Sciences , vol. 12, no. 11, pp. 5645, 2022, doi: 10.3390/app12115645

2022 doi
[13]

Image Fundus Classification System for Diabetic Retinopathy Stage Detection Using Hybrid CNN-DELM,

D. C. R. Novitasari et al., “Image Fundus Classification System for Diabetic Retinopathy Stage Detection Using Hybrid CNN-DELM,” Big Data Cogn. Comput. , vol. 6, no. 4, pp. 146, 2022, doi: 10.3390/bdcc6040146

2022 doi
[14]

Trans- fer Learning-Based Model for Diabetic Retinopathy Diagnosis Using Retinal Images,

M. K. Jabbar, J. Yan, H. Xu, Z. Ur Rehman, and A. Jabbar, “Trans- fer Learning-Based Model for Diabetic Retinopathy Diagnosis Using Retinal Images,” Brain Sciences , vol. 12, no. 5, p. 535, 2022, doi: 10.3390/brainsci12050535

2022 doi
[15]

Classifying Diabetic Retinopathy using CNN and Machine Learning,

C. Lahmar and A. Idri, “Classifying Diabetic Retinopathy using CNN and Machine Learning,” in Proceedings of the 15th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2022) - Volume 2: BIOIMAGING , pp. 52–62, 2022, doi: 10.5220/0010851500003123

2022 doi
[16]

https://www.openai.com/chatgpt (ac- cessed August 2024)

OpenAI, ”ChatGPT.” OpenAI. https://www.openai.com/chatgpt (ac- cessed August 2024)

2024

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

S. G. De Benedictis, G. Gargano, and G. Settembre, ”Enhanced MRI brain tumor detection and classification via topological data analy- sis and low-rank tensor decomposition,” Journal of Computational Mathematics and Data Science , vol. 13, 2024, Art. no. 100103. doi: 10.1016/j.jcmds.2024.100103

arXiv 2024

[2] [2]

Sikder, M

N. Sikder, M. S. Chowdhury, A. S. M. Arif, and A.-A. Nahid, ”Early blindness detection based on retinal images using ensemble learning,” in Proc. 2019 22nd Int. Conf. Comput. Inf. Technol. (ICCIT) , Dhaka, Bangladesh, 2019, pp. 1-6. doi: 10.1109/ICCIT48885.2019.9038439

work page arXiv 2019

[3] [3]

Breast Cancer Detection and Localizing the Mass Area Using Deep Learning,

M. M. Rahman et al., “Breast Cancer Detection and Localizing the Mass Area Using Deep Learning,” Big Data Cogn. Comput. , vol. 8, no. 80, 2024, doi: 10.3390/bdcc8070080

work page doi:10.3390/bdcc8070080 2024

[4] [4]

Two-Stage Selective Ensemble of CNN via Deep Tree Training for Medical Image Classification,

Y . Yang, Y . Hu, X. Zhang, and S. Wang, “Two-Stage Selective Ensemble of CNN via Deep Tree Training for Medical Image Classification,” IEEE Trans. Cybern. , vol. 52, no. 9, pp. 9194–9207, Sep. 2022, doi: 10.1109/TCYB.2021.3061147

arXiv 2022

[5] [5]

Feature fusion and Ensemble learning-based CNN model for mammographic image classification,

I. Ul Haq, H. Ali, H. Y . Wang, C. Lei, and H. Ali, “Feature fusion and Ensemble learning-based CNN model for mammographic image classification,” J. King Saud Univ. Comput. Inf. Sci. , vol. 34, no. 6, Part B, pp. 3310–3318, 2022, doi: 10.1016/j.jksuci.2022.03.023

work page doi:10.1016/j.jksuci.2022.03.023 2022

[6] [6]

Lung can- cer medical images classification using hybrid CNN-SVM,

A. Y . Saleh, C. K. Chin, V . Penshie, and H. Al-Absi, “Lung can- cer medical images classification using hybrid CNN-SVM,” Int. J. Adv. Intell. Informatics , vol. 7, no. 2, pp. 151–162, Jul. 2021, doi: 10.26555/ijain.v7i2.317

work page doi:10.26555/ijain.v7i2.317 2021

[7] [7]

Transfer Learning in Breast Cancer Diagnoses via Ultrasound Imaging,

G. Ayana, K. Dese, and S.-w. Choe, “Transfer Learning in Breast Cancer Diagnoses via Ultrasound Imaging,” Cancers, vol. 13, no. 738, 2021, doi: 10.3390/cancers13040738

work page doi:10.3390/cancers13040738 2021

[8] [8]

Towards a Better Understanding of Transfer Learning for Medical Imaging: A Case Study,

L. Alzubaidi et al., “Towards a Better Understanding of Transfer Learning for Medical Imaging: A Case Study,” Applied Sciences , vol. 10, no. 13, pp. 4523, 2020, doi: 10.3390/app10134523

work page doi:10.3390/app10134523 2020