Pith. sign in

REVIEW 5 major objections 5 minor 58 references

Bidirectional Brain Image Translation using Transfer Learning from Generic Pre-trained Models

T0 review · 5 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read A generic CycleGAN pretrained on flower photos, after light fine-tuning, outperforms 17 other non-medical checkpoints for bidirectional MR-CT brain synthesis and yields images radiologists rate nearly as real as ground truth.

desk verdict A useful checkpoint-scan experiment buried under an unsupported headline claim and an incoherent fine-tuning description. read the letter →

arxiv 2501.12488 v1 pith:ICQ4SPGW submitted 2025-01-21 eess.IV cs.CVq-bio.TO

classification eess.IVcs.CVq-bio.TO
keywords imagetranslationtransferlearningpre-trainedmodelsCycleGANbraintumormagneticresonanceimagingcomputedtomographymedicalsynthesis
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 tries to show that generic, non-medical pretrained CycleGAN checkpoints can be transferred to bidirectional brain MR-CT image synthesis, and that the choice of source images matters. Using 18 pretrained models covering artistic styles, animals, landscapes, photography, satellite maps, and urban scenes, the authors fine-tune each on 367 paired MR-CT brain slices and evaluate PSNR, SSIM, UQI, and VIF. The iphone2dslr flower checkpoint, trained on iPhone and DSLR flower photos, ranks first on all four metrics in both translation directions, with PSNR 30.98 for MR-to-CT and 34.36 for CT-to-MR. In a blinded perceptual study, two radiologists classified its synthetic images as real about 98% of the time, close to the ground-truth rate, and the model showed clean separation of MR and CT in latent space. If right, the result suggests that a cheap, publicly available non-medical checkpoint whose textures resemble brain tissue can reduce the data and compute needed for clinically plausible synthesis.

What carries the argument

The machinery is a CycleGAN backbone, an unsupervised generative network with two generators and two discriminators trained with adversarial loss plus cycle-consistency loss so a translated image can be mapped back to its source, combined with a transfer-learning protocol. Each pretrained checkpoint is loaded with its original weights; two fully connected layers, FC1 and FC2 with 256 neurons each and softmax activation, are added and randomly initialized; all original layers are frozen; and the model is trained for 200 epochs on the paired MR-CT dataset with lambda values of 9 to 11 depending on the source category and a learning rate that linearly decays after epoch 100. The frozen pretrained layers are meant to retain generic visual features while the added layers adapt the model to the MR and CT domains, and the latent-space analysis is offered as evidence that the best model separates the two modalities cleanly.

What would settle it

Train the same CycleGAN architecture with the same paired MR-CT data, splits, and hyperparameters from randomly initialized weights, and compare PSNR, SSIM, UQI, and VIF on the same test slices; if the from-scratch model matches or exceeds the flower checkpoint's PSNR of about 30.98 for MR-to-CT and 34.36 for CT-to-MR, the claimed transfer advantage is not real. A complementary check is to rerun the fine-tuning without the two added fully connected layers and with the pretrained layers unfrozen, to see whether the reported ranking survives changes to the adaptation layer.

Watch

Extended reading notes

Core claim

The central claim is that transfer learning from generic pretrained models is a viable and effective route for bidirectional MR-CT brain image translation, and that the iphone2dslr flower checkpoint is the best of the 18 generic models for the task. After fine-tuning all checkpoints under the same protocol, adding two 256-neuron fully connected layers with softmax, freezing the pretrained weights, and training for 200 epochs with category-dependent lambda, the flower model achieves the top score in all four evaluation scales in both directions, including PSNR 30.98 plus or minus 0.119 and SSIM 0.65 plus or minus 0.031 for MR-to-CT, and PSNR 34.36 plus or minus 0.072 and SSIM 0.83 plus or minus 0.022 for CT-to-MR. Radiologists rated its outputs 3.79 out of 4 for MR-to-CT and 3.69 out of 4 for CT-to-MR, with 97.91% and 97.7% of the images judged real, against 98.58% and 98.37% for ground truth. The paper attributes the advantage to the high quality and structural similarity of flower images to brain tissue, including petal patterns and convoluted surfaces that resemble cerebral structures, and argues that careful selection of representative training images is decisive for medical synthesis.

Load-bearing premise

The load-bearing premise is that adding two randomly initialized fully connected layers with softmax to a frozen generic CycleGAN checkpoint is genuine transfer learning; the paper gives no ablation of those added layers and no comparison to training the same CycleGAN from scratch, so the reported ranking could reflect initialization artifacts rather than transferred features.

Editorial extensions

If this is right

  • If the central claim holds, a checkpoint trained on non-medical images that resemble brain textures can replace a large medical training set as the starting point for MR-CT synthesis.
  • The iphone2dslr flower checkpoint becomes the default choice among the 18 tested generic models for this bidirectional brain-synthesis task.
  • Synthetic CT generated from MR could reduce unnecessary radiation exposure by supplying missing modalities without new scans.
  • The latent-space separation observed for the winning model offers a diagnostic signal for whether a pretrained model has actually adapted to the MR and CT domains.

Reading between the lines

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

  • A decisive control the paper leaves implicit is training the same CycleGAN from scratch on the same data; without it, the ranking is compatible with the added fully connected layers and training schedule doing most of the work.
  • The flower model's advantage may come from low-level texture statistics shared between petals and brain tissue; repeating the 18-model ranking on lung or cardiac CT-MR data would test whether the advantage transfers to other anatomies.
  • The clean MR-CT separation in the latent space suggests a cheap screening procedure: measure latent separability on a validation split to pick a pretrained checkpoint before committing to full fine-tuning.
  • The roughly 98% radiologist realism rate comes from two readers at one institution on one dataset, so it supports further study rather than immediate clinical deployment.
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 / 5 minor

Summary. The paper claims that transfer learning from 18 generic non-medical CycleGAN checkpoints can perform bidirectional MR-CT brain image translation, and reports that the iphone2dslr flower checkpoint is the best, with radiologists rating its outputs nearly as realistic as ground truth. The authors describe fine-tuning by freezing pre-existing layers and adding two fully connected layers with Softmax, report PSNR, SSIM, UQI, and VIF results for all 18 checkpoints, and include a radiologist perceptual study and latent-space visualization. The central claim is that a generic flower-photo model outperforms all other non-medical checkpoints and approaches clinical realism.

Significance. If the central claim held, the paper would provide a striking data point about which properties of pre-trained image-translation models transfer to medical modalities and would support a practical recipe for low-data MR-CT synthesis. The study is also useful as a broad benchmark of 18 public CycleGAN checkpoints on a fixed paired brain dataset, and the radiologist study is a commendable attempt to go beyond pixel metrics. However, the significance is undermined by the paper's own tables, which contradict the headline claim, and by an insufficiently specified fine-tuning procedure that prevents the reader from verifying that any transfer actually occurred.

major comments (5)
  1. [§III.B, Algorithm 1] The fine-tuning protocol is not architecturally coherent. CycleGAN's generator is fully convolutional (c7s1-64, d128, d256, residual blocks, u128, u64, c7s1-3) and outputs a 256×256×3 image. Placing two randomly initialized fully connected layers with Softmax after this generator, while freezing all original layers, cannot produce a 256×256×3 output from the image-sized generator output unless the FC layers are preceded by a flattening and followed by a reshaping that is not specified. As written, Eqs. (1)-(3) are defined over images and cannot train such a modified generator. The paper gives no tensor shapes, no model summary, and no code, so the described mechanism cannot be reproduced or verified.
  2. [Table III] The stated conclusion that iphone2dslr flower achieves 'top scores in all four evaluation scales' is contradicted by Table III. In the MR-CT direction, iphone2dslr flower has UQI = 0.00 and VIF = 0.01, while winter2summer yosemite has UQI = 0.04 and zebra2horse has VIF = 0.05; summer2winter yosemite also has higher SSIM. In the CT-MR direction, several models have higher VIF (e.g., winter2summer yosemite at 0.14) and higher UQI (e.g., winter2summer yosemite at 0.04). The paper itself states in the text that summer2winter yosemite has the highest SSIM and UQI and zebra2horse has the highest VIF in the CT-MR table. The headline claim is therefore not supported by the reported data.
  3. [§IV.A, §IV.C] No from-scratch CycleGAN baseline is included. The paper's central claim is that pre-trained generic checkpoints transfer useful features for MR-CT synthesis, but without training the same architecture from random initialization on the same data, the reported ranking of 18 checkpoints could reflect initialization artifacts, optimization path differences, or even the added FC layers rather than transfer of learned representations. This missing baseline is load-bearing for the conclusion that transfer learning, rather than the fine-tuning procedure itself, is responsible for the results.
  4. [§V, Discussion] The explanation that iphone2dslr flower succeeds because flower images resemble brain structures is post hoc. The paper does not report any pre-registered hypothesis or any quantitative measure of visual similarity between the flower dataset and brain images; it presents the resemblance as an observation made after identifying the winner. As stated, this is a rationalization rather than a tested explanation, and it cannot be distinguished from overfitting to the specific test split or from random checkpoint variation.
  5. [§IV.D, Table V] The perceptual study is reported only for the best model, selected using the same test data and metrics reported in Tables III-IV. Since the selection was made on the evaluation set, the radiologist realism scores are not an independent validation of the model's quality; they confirm only that the already-selected checkpoint produces images that radiologists find realistic. The paper does not report whether the same radiologists rated outputs of other checkpoints or a from-scratch baseline, so the perceptual evidence cannot support the claim that transfer learning from this checkpoint is uniquely beneficial.
minor comments (5)
  1. [Abstract and §V] The abstract and discussion state that results provide 'compelling evidence' of exceptional performance, but the internal contradictions in Tables III-IV should be acknowledged and reconciled before making such strong claims.
  2. [Table III and IV] The table headers are misaligned: 'A VG STD' appears to be a corrupted rendering of 'AVG', and the columns under each metric are not clearly labeled in the manuscript text. This makes the tables harder to read than necessary.
  3. [§IV.B, Eq. (4)] The PSNR equation is written as PSNR = 20 log10(R²/MSE), which is dimensionally incorrect; the standard definition is 10 log10(R²/MSE). This appears to be a typographical error, but it should be corrected.
  4. [Throughout] There are numerous typos and incomplete sentences, including 'Translation from MR–CT brain tumors presented' at the start of Section IV and 'pre-trained model' inconsistently capitalized in Section III.C. The paper would benefit from careful proofreading.
  5. [References] Several references appear to be cited inconsistently: [15] is invoked for multiple different works, and reference [43] (VIF) is not discussed in the text with the same depth as the other metrics. The reference list should be checked for accuracy and completeness.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported ranking is an empirical model-selection outcome evaluated against ground truth, not a quantity fitted into the evaluation by construction.

full rationale

The paper's central claim is an empirical ranking of 18 pretrained CycleGAN checkpoints for bidirectional MR-CT synthesis. The evaluation uses standard full-reference metrics (PSNR, SSIM, UQI, VIF) against ground-truth test images and a radiologist perceptual study; these are external criteria, not quantities reconstructed from the method's own definitions. The fine-tuning description in Section III.B is architecturally problematic (adding two FC+Softmax layers to a fully convolutional generator while freezing all other layers cannot produce the 256x256x3 image output required by the CycleGAN losses in Eqs. (1)-(3)), and the Discussion's statement that iphone2dslr flower achieved 'top scores in all four evaluation scales' is contradicted by Table III, where that model has UQI=0.00 and VIF=0.01 in the MR-CT direction. These are correctness and internal-consistency defects, not circularity: no reported metric is defined in terms of the model choice or reduced to the fitted hyperparameters by construction. The 'flower images resemble brain tissue' explanation in Section V is a post hoc interpretation offered after observing the winner, rather than a pre-registered prediction, but an after-the-fact explanation is not a circular derivation. The paper itself also acknowledges in Section V that the iphone2dslr flower model 'is not directly designed for medical image translation,' which is a limitation but not a circular step. Self-citations, including [2] as the dataset source and [1] for metric formulas, are not load-bearing: the dataset is a published resource and the metrics are standard external benchmarks. The central comparison across 18 publicly available checkpoints is self-contained empirical work, so no circular step is exhibited.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central comparison rests on a small self-sourced dataset, tuned hyperparameters, and a post hoc explanation of the winning model. No new entities are postulated. The main unstated premises are that the fine-tuning scheme transfers features at all, that the dataset is representative, and that the flower-brain similarity explanation is causal rather than descriptive.

free parameters (2)
  • lambda (cycle consistency weight) = 9 (artistic), 10 (animals, natural, photography), 11 (satellite, urban)
    Set per category based on experiments; Section III.B states that results are sensitive to the lambda value.
  • training hyperparameters = batch=2, FC1=256, FC2=256, epochs=200, learning rate=0.001
    Described as 'carefully selected based on the outcomes of our experiments' in Section III.B, with no separate validation split or sensitivity analysis reported.
assumptions (3)
  • domain assumption The 367 paired MR-CT brain images from 18 patients are representative of brain imaging for this task.
    Section IV.A describes the dataset from prior work [2] and uses only a single small cohort; no external validation or multi-site data is used.
  • ad hoc to paper Freezing pretrained layers and adding two fully connected layers preserves useful representations for MR-CT synthesis.
    Section III.B describes this fine-tuning procedure without architectural justification, ablation, or comparison to a from-scratch baseline.
  • ad hoc to paper Visual similarity between flower images and brain structures explains the ranking of the pre-trained models.
    The Discussion (Section V) invokes this resemblance after the fact to explain why iphone2dslr flower won; it is not tested as a prior hypothesis.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Bidirectional Brain Image Translation using Transfer Learning from Generic Pre-trained Models." pith.science (2026). https://pith.science/paper/ICQ4SPGW

@misc{pith2026250112488,
  author       = {Pith},
  title        = {Pith review of: Bidirectional Brain Image Translation using Transfer Learning from Generic Pre-trained Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ICQ4SPGW}},
  note         = {Machine review of arXiv:2501.12488}
}
read the original abstract

Brain imaging plays a crucial role in the diagnosis and treatment of various neurological disorders, providing valuable insights into the structure and function of the brain. Techniques such as magnetic resonance imaging (MRI) and computed tomography (CT) enable non-invasive visualization of the brain, aiding in the understanding of brain anatomy, abnormalities, and functional connectivity. However, cost and radiation dose may limit the acquisition of specific image modalities, so medical image synthesis can be used to generate required medical images without actual addition. In the medical domain, where obtaining labeled medical images is labor-intensive and expensive, addressing data scarcity is a major challenge. Recent studies propose using transfer learning to overcome this issue. This involves adapting pre-trained CycleGAN models, initially trained on non-medical data, to generate realistic medical images. In this work, transfer learning was applied to the task of MR-CT image translation and vice versa using 18 pre-trained non-medical models, and the models were fine-tuned to have the best result. The models' performance was evaluated using four widely used image quality metrics: Peak-signal-to-noise-ratio, Structural Similarity Index, Universal Quality Index, and Visual Information Fidelity. Quantitative evaluation and qualitative perceptual analysis by radiologists demonstrate the potential of transfer learning in medical imaging and the effectiveness of the generic pre-trained model. The results provide compelling evidence of the model's exceptional performance, which can be attributed to the high quality and similarity of the training images to actual human brain images. These results underscore the significance of carefully selecting appropriate and representative training images to optimize performance in brain image analysis tasks.

Figures

Figures reproduced from arXiv: 2501.12488 by the authors.

Figure 1
Figure 1. Transfer learning from generic pre-trained models. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. CycleGAN Generators and Discriminators [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Examples of generic pre-trained models [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: VIF components C. Comparison of a pre-trained model In image synthesis tasks, the evaluation of generated im￾ages is crucial to assessing the quality of the model’s out￾put. The commonly used PSNR metric, which calculates the mean square error between the original and …
Figure 5
Figure 5. Figure 5: Output results of Magnetic Resonance to Computed Tomography (MR-CT) image translation from different generic [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Output results of Computed Tomography to Magnetic Resonance (CT-MR) image translation from different generic [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: Latent space visualization using different pre-trained generic models for MR-CT and CT-MR image translation. [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: Sample images from generic iphone2dslr flower pre￾trained model. mentioning that the iphone2dslr flower model is a conditional generative model that can generate realistic flower images from a given input image, it is not directly designed for medical image translation…
Figure 9
Figure 9. Figure 9: Physiological characteristic similarities between [PITH_FULL_IMAGE:figures/full_fig_p017_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 56 canonical work pages

  1. [1]

    The effect of loss function on con- ditional generative adversarial networks

    Alaa Abu-Srhan, Mohammad AM Abushariah, and Omar S Al-Kadi. “The effect of loss function on con- ditional generative adversarial networks”. In: Journal of King Saud University-Computer and Information Sciences 34.9 (2022), pp. 6977–6988

  2. [2]

    Paired-unpaired Unsupervised Attention Guided GAN with transfer learning for bidi- rectional brain MR-CT synthesis

    Alaa Abu-Srhan et al. “Paired-unpaired Unsupervised Attention Guided GAN with transfer learning for bidi- rectional brain MR-CT synthesis”. In: Computers in Biology and Medicine 136 (2021), p. 104763

  3. [3]

    Optimal and Efficient Deep Learning Model for Brain Tumor Magnetic Resonance Imaging Classification and Analysis

    Manar Ahmed Hamza et al. “Optimal and Efficient Deep Learning Model for Brain Tumor Magnetic Resonance Imaging Classification and Analysis”. In: Applied Sci- ences 12.15 (2022), p. 7953

  4. [4]

    Transfer learning for automatic brain tumor classification using MRI images

    Mohamed Arbane et al. “Transfer learning for automatic brain tumor classification using MRI images”. In: 2020 2nd International Workshop on Human-Centric Smart Environments for Health and Well-being (IHSH) . IEEE. 2021, pp. 210–214

  5. [5]

    Adjuvant procarbazine, lomustine and vincristine chemotherapy in newly diag- nosed anaplastic oligodendroglioma: long-term follow- up of EORTC brain tumor group study 26951

    Martin J van den Bent et al. “Adjuvant procarbazine, lomustine and vincristine chemotherapy in newly diag- nosed anaplastic oligodendroglioma: long-term follow- up of EORTC brain tumor group study 26951”. In: Journal of clinical oncology 31.3 (2013), pp. 344–350

  6. [6]

    Exploring the Power of Generative Deep Learning for Image-to-Image Translation and MRI Reconstruction: A Cross-Domain Review

    Yuda Bi. “Exploring the Power of Generative Deep Learning for Image-to-Image Translation and MRI Reconstruction: A Cross-Domain Review”. In: arXiv preprint arXiv:2303.09012 (2023)

  7. [7]

    Transfer learning using convolutional neural network architectures for brain tumor classification from MRI images

    Rayene Chelghoum et al. “Transfer learning using convolutional neural network architectures for brain tumor classification from MRI images”. In: Artificial Intelligence Applications and Innovations: 16th IFIP WG 12.5 International Conference, AIAI 2020, Neos Marmaras, Greece, June 5–7, 2020, Proceedings, Part I 16. Springer. 2020, pp. 189–200

  8. [8]

    Not-so-supervised: a survey of semi- supervised, multi-instance, and transfer learning in med- ical image analysis

    Veronika Cheplygina, Marleen de Bruijne, and Josien PW Pluim. “Not-so-supervised: a survey of semi- supervised, multi-instance, and transfer learning in med- ical image analysis”. In: Medical image analysis 54 (2019), pp. 280–296

Show all 58 references
  1. [9]

    Detection of tumors on brain MRI images using the hybrid con- volutional neural network architecture

    Ahmet Cinar and Muhammed Yildirim. “Detection of tumors on brain MRI images using the hybrid con- volutional neural network architecture”. In: Medical hypotheses 139 (2020), p. 109684

  2. [10]

    The cityscapes dataset for se- mantic urban scene understanding

    Marius Cordts et al. “The cityscapes dataset for se- mantic urban scene understanding”. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2016, pp. 3213–3223

  3. [11]

    Automated categorization of brain tumor from MRI using cnn features and svm

    S Deepak and PM Ameer. “Automated categorization of brain tumor from MRI using cnn features and svm”. In: Journal of Ambient Intelligence and Humanized Computing 12 (2021), pp. 8357–8369

  4. [12]

    Imagenet: A large-scale hierarchical image database

    Jia Deng et al. “Imagenet: A large-scale hierarchical image database”. In: 2009 IEEE conference on computer vision and pattern recognition. Ieee. 2009, pp. 248–255

  5. [13]

    Biomedical texture analysis: fundamentals, tools and challenges

    Adrien Depeursinge, Omar S Al-Kadi, and J Ross Mitchell. Biomedical texture analysis: fundamentals, tools and challenges . Academic Press, 2017

  6. [14]

    Computational Fractal-Based Analysis of Brain Tumor Microvascular Networks

    Antonio Di Ieva and Omar S Al-Kadi. “Computational Fractal-Based Analysis of Brain Tumor Microvascular Networks”. In: The Fractal Geometry of the Brain . Springer, 2024, pp. 525–544

  7. [15]

    Deepfake Image Generation for Improved Brain Tumor Segmentation

    Roa’ Al-Emaryeen et al. “Deepfake Image Generation for Improved Brain Tumor Segmentation”. In: 2023 18 IEEE Jordan International Joint Conference on Electri- cal Engineering and Information Technology (JEEIT) . IEEE. 2023, pp. 6–11

  8. [16]

    On handling the negative transfer and imbalanced distributions in multiple source transfer learning

    Liang Ge et al. “On handling the negative transfer and imbalanced distributions in multiple source transfer learning”. In: Statistical Analysis and Data Mining: The ASA Data Science Journal 7.4 (2014), pp. 254–271

  9. [17]

    Lesion-preserving unpaired image-to-image translation between MRI and CT from ischemic stroke patients

    Alejandro Gutierrez et al. “Lesion-preserving unpaired image-to-image translation between MRI and CT from ischemic stroke patients”. In: International Journal of Computer Assisted Radiology and Surgery (2023), pp. 1–10

  10. [18]

    A transfer learning–based active learning framework for brain tumor classification

    Ruqian Hao et al. “A transfer learning–based active learning framework for brain tumor classification”. In: Frontiers in Artificial Intelligence 4 (2021), p. 635766

  11. [19]

    Image-to-image translation with conditional adversarial networks

    Phillip Isola et al. “Image-to-image translation with conditional adversarial networks”. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2017, pp. 1125–1134

  12. [20]

    A gabor filter texture analysis approach for histopathological brain tumor subtype discrimination

    Omar S Al-Kadi. “A gabor filter texture analysis approach for histopathological brain tumor subtype discrimination”. In: arXiv preprint arXiv:1704.05122 (2017)

  13. [21]

    Fractals for biomedical texture anal- ysis

    Omar S Al-Kadi. “Fractals for biomedical texture anal- ysis”. In: Biomedical texture analysis . Elsevier, 2017, pp. 131–161

  14. [22]

    Fractal-Based Analysis of Histological Features of Brain Tumors

    Omar S Al-Kadi and Antonio Di Ieva. “Fractal-Based Analysis of Histological Features of Brain Tumors”. In: The Fractal Geometry of the Brain . Springer, 2024, pp. 501–524

  15. [23]

    Reviews in cancer imaging and image-directed interventions

    Omar S Al-Kadi and Oliver Diaz. “Reviews in cancer imaging and image-directed interventions”. In: Frontiers in Oncology 13 (2023), p. 1183302

  16. [24]

    Texture analysis of aggressive and nonaggressive lung tumor CE CT im- ages

    Omar S Al-Kadi, D Watson, et al. “Texture analysis of aggressive and nonaggressive lung tumor CE CT im- ages”. In: IEEE transactions on biomedical engineering 55.7 (2008), pp. 1822–1830

  17. [25]

    Empowering brain cancer diagnosis: harnessing artificial intelligence for advanced imaging insights

    Omar S. Al-Kadi et al. “Empowering brain cancer diagnosis: harnessing artificial intelligence for advanced imaging insights”. In: Reviews in the Neurosciences35.4 (2024), pp. 399–419

  18. [26]

    Tumour grading and discrimina- tion based on class assignment and quantitative texture analysis techniques

    Omar Sultan Al-Kadi. “Tumour grading and discrimina- tion based on class assignment and quantitative texture analysis techniques”. PhD thesis. University of Sussex, 2010

  19. [27]

    Deep convo- lutional neural networks with transfer learning for au- tomated brain image classification

    Taranjit Kaur and Tapan Kumar Gandhi. “Deep convo- lutional neural networks with transfer learning for au- tomated brain image classification”. In: Machine Vision and Applications 31.3 (2020), p. 20

  20. [28]

    A novel ap- proach to predict brain cancerous tumor using transfer learning

    Mohammad Monirujjaman Khan et al. “A novel ap- proach to predict brain cancerous tumor using transfer learning”. In: Computational and Mathematical Meth- ods in Medicine 2022 (2022)

  21. [29]

    Latent space model for higher-order networks and generalized tensor decomposition

    Zhongyuan Lyu, Dong Xia, and Yuan Zhang. “Latent space model for higher-order networks and generalized tensor decomposition”. In: Journal of Computational and Graphical Statistics (2023), pp. 1–17

  22. [30]

    Video Quality Assessment with Texture Information Fusion for Streaming Appli- cations

    Vignesh V Menon et al. “Video Quality Assessment with Texture Information Fusion for Streaming Appli- cations”. In: arXiv preprint arXiv:2302.14465 (2023)

  23. [31]

    Data augmentation and trans- fer learning for limited dataset ship classification

    Mario Milicevic et al. “Data augmentation and trans- fer learning for limited dataset ship classification”. In: WSEAS Trans. Syst. Control 13.1 (2018), pp. 460–465

  24. [32]

    Erratum: Medical Image Synthesis with Deep Convolutional Adversarial Networks (IEEE Trans. Biomed. Eng.(2018) 65: 12 (2720-2730

    Dong Nie et al. “Erratum: Medical Image Synthesis with Deep Convolutional Adversarial Networks (IEEE Trans. Biomed. Eng.(2018) 65: 12 (2720-2730”. In: IEEE Transactions on Biomedical Engineering 67.9 (2020), p. 2706

  25. [33]

    Medical image synthesis with context- aware generative adversarial networks

    Dong Nie et al. “Medical image synthesis with context- aware generative adversarial networks”. In: Medical Image Computing and Computer Assisted Intervention- MICCAI 2017: 20th International Conference, Quebec City, QC, Canada, September 11-13, 2017, Proceedings, Part III 20. ...

  26. [34]

    Con- ceptual and Hierarchical Latent Space Decomposition for Face Editing

    Savas Ozkan, Mete Ozay, and Tom Robinson. “Con- ceptual and Hierarchical Latent Space Decomposition for Face Editing”. In: Proceedings of the IEEE/CVF International Conference on Computer Vision . 2023, pp. 7248–7257

  27. [35]

    A survey on transfer learning

    Sinno Jialin Pan and Qiang Yang. “A survey on transfer learning”. In: IEEE Transactions on Knowledge and data engineering 22.10 (2010), pp. 1345–1359

  28. [36]

    A deep learning-based framework for automatic brain tumors classification using transfer learning

    Arshia Rehman et al. “A deep learning-based framework for automatic brain tumors classification using transfer learning”. In: Circuits, Systems, and Signal Processing 39 (2020), pp. 757–775

  29. [37]

    CyTran: A Cycle-Consistent Transformer with Multi-Level Consistency for Non- Contrast to Contrast CT Translation

    Nicolae-clin Ristea et al. “CyTran: A Cycle-Consistent Transformer with Multi-Level Consistency for Non- Contrast to Contrast CT Translation”. In: Neurocomput- ing (2023)

  30. [38]

    MR to CT registration of brains using image synthesis

    Snehashis Roy et al. “MR to CT registration of brains using image synthesis”. In: Medical Imaging 2014: Image Processing. V ol. 9034. SPIE. 2014, pp. 307–314

  31. [39]

    Data augmentation using generative adversarial networks (CycleGAN) to improve generaliz- ability in CT segmentation tasks

    Veit Sandfort et al. “Data augmentation using generative adversarial networks (CycleGAN) to improve generaliz- ability in CT segmentation tasks”. In: Scientific reports 9.1 (2019), p. 16884

  32. [40]

    Predictive modeling of brain tumor: a deep learning approach

    Priyansh Saxena, Akshat Maheshwari, and Saumil Maheshwari. “Predictive modeling of brain tumor: a deep learning approach”. In: Innovations in Computa- tional Intelligence and Computer Vision: Proceedings of ICICV 2020. Springer, 2020, pp. 275–285

  33. [41]

    PSNR vs SSIM: imperceptibility quality assessment for image steganog- raphy

    De Rosal Igantius Moses Setiadi. “PSNR vs SSIM: imperceptibility quality assessment for image steganog- raphy”. In: Multimedia Tools and Applications 80.6 (2021), pp. 8423–8444

  34. [42]

    Performance comparison of different pre- trained deep learning models in classifying brain MRI images

    Onur Sevli. “Performance comparison of different pre- trained deep learning models in classifying brain MRI images”. In: Acta Infologica 5.1 (2021), pp. 141–154

  35. [43]

    A visual infor- mation fidelity approach to video quality assessment

    Hamid R Sheikh and Alan C Bovik. “A visual infor- mation fidelity approach to video quality assessment”. In: The first international workshop on video processing and quality metrics for consumer electronics . V ol. 7. 2. sn. 2005, pp. 2117–2128. 19

  36. [45]

    Brain tumor classification for MR images using transfer learning and fine-tuning

    Zar Nawab Khan Swati et al. “Brain tumor classification for MR images using transfer learning and fine-tuning”. In: Computerized Medical Imaging and Graphics 75 (2019), pp. 34–46

  37. [46]

    Perfor- mance enhancement of MRI-based brain tumor classi- fication using suitable segmentation method and deep learning-based ensemble algorithm

    Gopal S Tandel, Ashish Tiwari, and OG Kakde. “Perfor- mance enhancement of MRI-based brain tumor classi- fication using suitable segmentation method and deep learning-based ensemble algorithm”. In: Biomedical Signal Processing and Control 78 (2022), p. 104018

  38. [47]

    Deep transfer learning for brain tumor detection based on MRI images

    Kasi Tenghongsakul et al. “Deep transfer learning for brain tumor detection based on MRI images”. In: Jour- nal of Physics: Conference Series . V ol. 2497. 1. IOP Publishing. 2023, p. 012015

  39. [48]

    Uvcgan: Unet vision trans- former cycle-consistent gan for unpaired image-to- image translation

    Dmitrii Torbunov et al. “Uvcgan: Unet vision trans- former cycle-consistent gan for unpaired image-to- image translation”. In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision . 2023, pp. 702–712

  40. [49]

    Spatial pattern tem- plates for recognition of objects with regular struc- ture

    Radim Tyle ˇcek and Radim ˇS´ara. “Spatial pattern tem- plates for recognition of objects with regular struc- ture”. In: Pattern Recognition: 35th German Confer- ence, GCPR 2013, Saarbr ¨ucken, Germany, September 3-6, 2013. Proceedings 35. Springer. 2013, pp. 364–374

  41. [50]

    Transfer learning in mag- netic resonance brain imaging: A systematic review

    Juan Miguel Valverde et al. “Transfer learning in mag- netic resonance brain imaging: A systematic review”. In: Journal of imaging 7.4 (2021), p. 66

  42. [51]

    Dc-cyclegan: bidirectional ct-to-mr synthesis from unpaired data

    Jiayuan Wang, QM Jonathan Wu, and Farhad Pour- panah. “Dc-cyclegan: bidirectional ct-to-mr synthesis from unpaired data”. In: Computerized Medical Imaging and Graphics (2023), p. 102249

  43. [52]

    A universal image quality index

    Zhou Wang and Alan C Bovik. “A universal image quality index”. In: IEEE signal processing letters 9.3 (2002), pp. 81–84

  44. [53]

    Characterizing and avoiding negative transfer

    Zirui Wang et al. “Characterizing and avoiding negative transfer”. In: Proceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition . 2019, pp. 11293–11302

  45. [54]

    A La- tent Space of Stochastic Diffusion Models for Zero- Shot Image Editing and Guidance

    Chen Henry Wu and Fernando De la Torre. “A La- tent Space of Stochastic Diffusion Models for Zero- Shot Image Editing and Guidance”. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. 2023, pp. 7378–7387

  46. [55]

    Glioma grading on conventional MR images: a deep learning study with transfer learning

    Yang Yang et al. “Glioma grading on conventional MR images: a deep learning study with transfer learning”. In: Frontiers in neuroscience 12 (2018), p. 804

  47. [56]

    Seismic impedance inversion based on geophysical-guided cycle-consistent generative adversarial networks

    Haihang Zhang et al. “Seismic impedance inversion based on geophysical-guided cycle-consistent generative adversarial networks”. In: Journal of Petroleum Science and Engineering 218 (2022), p. 111003

  48. [57]

    Unsupervised conditional consen- sus adversarial network for brain disease identification with structural MRI

    Jing Zhang et al. “Unsupervised conditional consen- sus adversarial network for brain disease identification with structural MRI”. In: Machine Learning in Medical Imaging: 10th International Workshop, MLMI 2019, Held in Conjunction with MICCAI 2019, Shenzhen, China, October 13...

  49. [58]

    Trans- lating and segmenting multimodal medical volumes with cycle-and shape-consistency generative adversar- ial network

    Zizhao Zhang, Lin Yang, and Yefeng Zheng. “Trans- lating and segmenting multimodal medical volumes with cycle-and shape-consistency generative adversar- ial network”. In: Proceedings of the IEEE conference on computer vision and Pattern Recognition . 2018, pp. 9242–9251

  50. [59]

    Unpaired image-to-image transla- tion using cycle-consistent adversarial networks

    Jun-Yan Zhu et al. “Unpaired image-to-image transla- tion using cycle-consistent adversarial networks”. In: Proceedings of the IEEE international conference on computer vision. 2017, pp. 2223–2232

Pith tools

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