Pith. sign in

REVIEW 3 major objections 1 minor 45 references

Mamba models chain visual image probabilities into tabular clinical data for multimodal cancer classification.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

Mamba-based visual and tabular models with mixed fusion achieve competitive balanced accuracy and higher recall than prior methods on skin and oral cancer datasets while enabling SHAP interpretability.

T0 review reviewed 2026-06-26 challenge →

load-bearing objection The paper's main move is a two-stage Mixed Fusion that feeds visual Mamba class probabilities into a tabular Mamba, gaining recall on one small dataset while trailing transformers on balanced accuracy for the other. the 3 major comments →

arxiv 2606.20738 v1 pith:W5XPO35Y submitted 2026-06-17 cs.CV cs.AI

An approach with Visual and Tabular Mamba to multimodal medical data using Mixed Fusion

Matheus B. Rocha , Gustavo B. Dettogni , Renato A. Krohling This is my paper
classification cs.CV cs.AI
keywords Mambamultimodal fusionmedical image classificationskin lesionsoral cancerSHAP interpretabilitystate space models
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

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 develops a Mixed Fusion architecture that runs a visual Mamba on lesion images to produce class probabilities, then feeds those probabilities together with clinical or sociodemographic variables into a separate tabular Mamba for the final diagnosis. Experiments on the PAD-UFES-20 skin-lesion dataset and the NDB-UFES oral-cancer dataset show competitive balanced accuracy with Transformers on the first set and better results on the second, together with clear gains in recall and the ability to apply SHAP explanations. The central aim is to demonstrate that this decoupled Mamba pipeline offers a practical alternative for multimodal medical tasks when high sensitivity matters.

Core claim

A visual Mamba classifies the input image and emits a probability vector over target classes; a tabular Mamba then receives that vector concatenated with clinical variables and outputs the final diagnosis, achieving superior recall on NDB-UFES while enabling direct SHAP attribution on the tabular stage.

What carries the argument

Mixed Fusion architecture that passes only the visual Mamba's class-probability vector to the tabular Mamba without joint end-to-end training or cross-attention layers.

Load-bearing premise

The class-probability vector from the visual Mamba is a sufficient summary of the image that the tabular Mamba can combine with clinical variables without loss of essential diagnostic information.

What would settle it

An experiment that replaces the probability vector with full visual features or trains the two Mambas jointly and measures a statistically significant drop in recall on NDB-UFES would falsify the claim that the decoupled probability summary suffices.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • SHAP can be applied directly to the tabular stage to attribute importance to both clinical variables and the visual probabilities.
  • The approach yields higher recall than some Transformer baselines on oral-cancer data while remaining competitive on skin-lesion data.
  • No additional cross-attention modules are required between modalities.
  • The pipeline separates image classification from tabular fusion, allowing each Mamba to be trained or updated independently.

Where Pith is reading between the lines

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

  • The probability-only interface may lower memory and compute costs relative to full multimodal attention models.
  • The same two-stage pattern could be tested on other paired image-tabular medical tasks such as radiology reports with patient metadata.
  • If the probability summary discards fine-grained image cues, richer intermediate representations could be passed instead.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 1 minor

Summary. The manuscript proposes a Mixed Fusion architecture that employs a Visual Mamba to classify lesion images and produce class probability vectors, which are then concatenated with tabular clinical or sociodemographic variables and processed by a separate Tabular Mamba to yield the final multimodal diagnosis. Experiments are reported on the PAD-UFES-20 skin-lesion dataset and the NDB-UFES oral-cancer dataset, claiming competitive or superior recall relative to Transformer baselines together with improved interpretability via SHAP.

Significance. If the reported recall improvements prove robust, the work would establish Mamba-based state-space models as a computationally lighter and more interpretable alternative to attention-based multimodal fusion in medical classification tasks where sensitivity matters. The explicit use of SHAP on the tabular stage is a concrete strength that supports the interpretability claim.

major comments (3)
  1. [Mixed Fusion architecture] Mixed Fusion architecture description: the two-stage pipeline treats the low-dimensional class-probability vector emitted by the visual Mamba as a sufficient and lossless summary of the image; without an ablation that compares this design to joint end-to-end training or to feature-level cross-attention, the assumption that no diagnostically relevant image–tabular interactions are lost remains untested and directly bears on the multimodal claim.
  2. [Experimental results] Experimental evaluation (PAD-UFES-20 and NDB-UFES results): performance numbers are presented without training details, hyper-parameter search protocol, number of independent runs, cross-validation scheme, or statistical significance tests, rendering it impossible to judge whether the reported recall gains are stable or reproducible.
  3. [Experimental results] Baseline comparison: the claim of suitability as an alternative to Transformer-based approaches rests on the reported metrics, yet no implementation details or ablation isolating the contribution of the fusion step versus the Mamba backbone itself are supplied.
minor comments (1)
  1. [Methods] The abstract and methods would benefit from explicit equations defining the visual-Mamba output probability vector and its concatenation with tabular inputs before the tabular Mamba.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. We address each major point below, indicating the revisions we will incorporate to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Mixed Fusion architecture] Mixed Fusion architecture description: the two-stage pipeline treats the low-dimensional class-probability vector emitted by the visual Mamba as a sufficient and lossless summary of the image; without an ablation that compares this design to joint end-to-end training or to feature-level cross-attention, the assumption that no diagnostically relevant image–tabular interactions are lost remains untested and directly bears on the multimodal claim.

    Authors: The Mixed Fusion design was chosen to enable direct application of SHAP to the tabular Mamba stage, which receives the class-probability vector together with clinical variables; this yields the interpretability benefit highlighted in the manuscript. We acknowledge that the two-stage separation assumes the probability vector is a sufficient summary and that an ablation against joint end-to-end training or feature-level cross-attention would be required to fully test whether diagnostically relevant interactions are lost. We will add such an ablation study in the revised version. revision: yes

  2. Referee: [Experimental results] Experimental evaluation (PAD-UFES-20 and NDB-UFES results): performance numbers are presented without training details, hyper-parameter search protocol, number of independent runs, cross-validation scheme, or statistical significance tests, rendering it impossible to judge whether the reported recall gains are stable or reproducible.

    Authors: We agree that the current experimental section omits these critical details. In the revision we will supply the full training protocol, hyper-parameter search procedure, number of independent runs, cross-validation scheme, and statistical significance tests so that the stability and reproducibility of the recall improvements can be properly evaluated. revision: yes

  3. Referee: [Experimental results] Baseline comparison: the claim of suitability as an alternative to Transformer-based approaches rests on the reported metrics, yet no implementation details or ablation isolating the contribution of the fusion step versus the Mamba backbone itself are supplied.

    Authors: We will expand the baseline section with complete implementation details for all compared models and will add an ablation that isolates the contribution of the fusion step from the Mamba backbone, thereby clarifying whether the observed performance stems primarily from the fusion mechanism or from the state-space model itself. revision: yes

Circularity Check

0 steps flagged

No circularity: purely empirical architecture evaluation with no derivations or self-referential predictions

full rationale

The manuscript proposes and evaluates a Mixed Fusion architecture that chains a visual Mamba (producing class probabilities from images) into a tabular Mamba (combining those probabilities with clinical variables). All reported results are obtained by training on the PAD-UFES-20 and NDB-UFES datasets and measuring balanced accuracy, recall, and other metrics against Transformer baselines. No equations, first-principles derivations, fitted parameters renamed as predictions, or uniqueness theorems appear. The interpretability claim rests on post-hoc SHAP application, which is external to the model training. The central claim therefore remains an empirical observation rather than a reduction to its own inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The approach inherits the Mamba state-space model from prior work and relies on standard supervised classification assumptions; no new entities or free parameters are introduced in the abstract.

axioms (1)
  • domain assumption Visual and tabular streams can be trained independently and joined only through class probabilities without loss of diagnostic information.
    This premise is required for the Mixed Fusion design to be valid; it is stated implicitly by the two-stage architecture description.

reviewed 2026-06-26 · how reviews work

0 comments
Cite this review

Pith. "Pith review of An approach with Visual and Tabular Mamba to multimodal medical data using Mixed Fusion." pith.science (2026). https://pith.science/paper/W5XPO35Y

@misc{pith2026260620738,
  author       = {Pith},
  title        = {Pith review of: An approach with Visual and Tabular Mamba to multimodal medical data using Mixed Fusion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/W5XPO35Y}},
  note         = {Machine review of arXiv:2606.20738}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

This article presents a complementary approach for integrating multimodal medical data in cancer classification, based on state space models represented by the Mamba architecture. To this end, a mixed multimodal fusion architecture, called Mixed Fusion, was employed and developed to enhance the interpretability of the decision-making process. The proposed approach explores two variants of Mamba: one dedicated to visual processing, responsible for classifying the lesion image and generating probabilities associated with the target classes, and another focused on tabular processing, which uses these probabilities together with clinical and/or sociodemographic data to produce the final diagnosis. The experiments were conducted on two medical datasets: PAD-UFES-20, composed of clinical images and information associated with skin lesions, and NDB-UFES, consisting of histopathological images and sociodemographic data related to oral cancer. The results indicate slightly lower performance in balanced accuracy, compared with Transformer-based approaches, on PAD-UFES-20, and superior performance on NDB-UFES. Additionally, substantial gains were observed in the recall metric. Furthermore, the adoption of the Mixed Fusion architecture enables the application of the Shapley Additive Explanations (SHAP) method, increasing the interpretability of the results. These findings indicate that Mamba-based models constitute a suitable alternative for multimodal classification in medical data, especially in scenarios in which sensitivity is a relevant requirement.

Figures

Figures reproduced from arXiv: 2606.20738 by Gustavo B. Dettogni, Matheus B. Rocha, Renato A. Krohling.

Figure 1
Figure 1. Figure 1: Mixed fusion architecture. The images are initially processed by a feature ex￾tractor and then passed through a classifier that generates class probabilities. These probabilities are combined with clinical data, resulting in a final classification step In contrast to conventional approaches, the Mixed Fusion architecture makes it possible to integrate the SHapley Additive exPlanations (SHAP) method [26], … view at source ↗
Figure 2
Figure 2. Figure 2: Examples of images present in the PAD-UFES-20 dataset. 4.2 NDB-UFES Dataset The second dataset analyzed in this study is NDB-UFES [6]. In contrast to PAD￾UFES-20, this dataset is composed of histopathological images of oral lesions as well as clinical and sociodemographic information from patients diagnosed with oral squamous cell carcinoma (OSCC), oral leukoplakia with dysplasia (LW/D), and oral leukoplak… view at source ↗
Figure 3
Figure 3. Figure 3: Examples of images present in the NDB-UFES dataset [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Feature impact on model output for classes on PAD-UFES-20 using VMamba/Mambular and NDB-UFES using MambaVision/MambAttention. (“background_father_POMERANIA”), which may be related to genetic, envi￾ronmental, or socioeconomic factors. For the NDB-UFES dataset, the SHAP analysis was performed considering the combination of the MambaVision internal classifier and the MambAttention external classifier. Figure … view at source ↗

discussion (0)

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

Reference graph

Works this paper leans on

45 extracted references · 7 canonical work pages · 3 internal anchors

  1. [1]

    In: Interna- tional Conference on Robot Intelligence Technology and Applications

    Abdul Rauf, A.R., Mohd Isa, W.H., Khairuddin, I.M., et al.: The classification of oral squamous cell carcinoma (OSCC) by means of transfer learning. In: Interna- tional Conference on Robot Intelligence Technology and Applications. pp. 386–391. Springer (2022)

    2022

  2. [2]

    In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining

    Akiba, T., Sano, S., Yanase, T., Ohta, T., Koyama, M.: Optuna: A next- generation hyperparameter optimization framework. In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. p. 2623–2631. Association for Computing Machinery, New York, NY, USA (2019)

    2019

  3. [3]

    medRxiv preprint medRxiv:2021.05.06.21256741 (2021)

    Amin, I., Zamir, H., Khan, F.F.: Histopathological image analysis for oral squa- mous cell carcinoma classification using concatenated deep learning models. medRxiv preprint medRxiv:2021.05.06.21256741 (2021)

    2021

  4. [4]

    Revista Brasileira de Epidemiologia18, 894–905 (2015)

    Andrade, J.O.M., Santos, C.A.d.S.T., Oliveira, M.C.: Associated factors with oral cancer: a study of case control in a population of the Brazil’s Northeast. Revista Brasileira de Epidemiologia18, 894–905 (2015)

    2015

  5. [5]

    Journal of Reliable Intelligent En- vironments11(1), 1 (Dec 2024)

    Assis, A., Dantas, J., Andrade, E.: The performance-interpretability trade-off: a comparative study of machine learning models. Journal of Reliable Intelligent En- vironments11(1), 1 (Dec 2024)

    2024

  6. [6]

    Data in Brief48, 109128 (2023)

    Ribeiro-de Assis, M.C.F., Soares, J.P., de Lima, L.M., de Barros, L.A.P., Grão- Velloso, T.R., Krohling, R.A., Camisasca, D.R.: NDB-UFES: an oral cancer and leukoplakia dataset composed of histopathological images and patient data. Data in Brief48, 109128 (2023)

    2023

  7. [7]

    A pilot study

    del Carmen Migueláñez-Medrán, B., Pozo-Kreilinger, J.J., Cebrián-Carretero, J.L., Martínez-García, M.Á., López-Sánchez, A.F.: Oral squamous cell carcinoma of tongue: Histological risk assessment. A pilot study. Medicina Oral, Patología Oral y Cirugía Bucal24(5), e603 (2019)

    2019

  8. [8]

    Skin Health and Disease3(6), e295 (2023)

    Congdon, N.M., Davis, C.M.: A systematic review of the frequency of features of the seven-point checklist in proven cutaneous melanoma: The importance of change. Skin Health and Disease3(6), e295 (2023)

    2023

  9. [9]

    In: Skin Cancer: Pathogenesis and Diagnosis, pp

    Das, T., Kumar, V., Prakash, A., Lynn, A.M.: Artificial intelligence in skin cancer: diagnosis and therapy. In: Skin Cancer: Pathogenesis and Diagnosis, pp. 143–171. Springer (2021)

    2021

  10. [10]

    European Journal of Dermatology31(6), 771–778 (2021)

    Duarte, A.F., Sousa-Pinto, B., Azevedo, L.F., Barros, A.M., Puig, S., Malvehy, J., Haneke, E., Correia, O.: Clinical ABCDE rule for early melanoma detection. European Journal of Dermatology31(6), 771–778 (2021)

    2021

  11. [11]

    Cancers17(21) (2025)

    Garcia, A., Zhou, J., Pinero-Crespo, G., Beachkofsky, T., Huang, X.: Clinical ap- plication of vision transformers for melanoma classification: A multi-dataset eval- uation study. Cancers17(21) (2025)

    2025

  12. [12]

    In: Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., Oh, A

    Gorishniy, Y., Rubachev, I., Babenko, A.: On embeddings for numerical features in tabular deep learning. In: Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., Oh, A. (eds.) Advances in Neural Information Processing Systems. vol. 35, pp. 24991–25004. Curran Associates, Inc. (2022)

    2022

  13. [13]

    British Journal of General Practice69(679), e112–e126 (2019)

    Grafton-Clarke, C., Chen, K.W., Wilcock, J.: Diagnosis and referral delays in pri- mary care for oral squamous cell cancer: a systematic review. British Journal of General Practice69(679), e112–e126 (2019)

    2019

  14. [14]

    Mamba: Linear-Time Sequence Modeling with Selective State Spaces

    Gu, A., Dao, T.: Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752 (2024)

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  15. [15]

    arXiv preprint arXiv:2407.08083 (2025) 14 M

    Hatamizadeh, A., Kautz, J.: MambaVision: A Hybrid Mamba-Transformer Vision Backbone. arXiv preprint arXiv:2407.08083 (2025) 14 M. Rocha, G. Dettogni and R. Krohling

    work page arXiv 2025

  16. [16]

    Journal of Dental Research99(3), 241–248 (2020)

    Ilhan, B., Lin, K., Guneri, P., et al: Improving oral cancer outcomes with imaging and artificial intelligence. Journal of Dental Research99(3), 241–248 (2020)

    2020

  17. [17]

    Quantitative Imaging in Medicine and Surgery14(5), 3501–3518 (2024)

    Jia, H., Zhang, J., Ma, K., Qiao, X., Ren, L., Shi, X.: Application of convolutional neural networks in medical images: a bibliometric analysis. Quantitative Imaging in Medicine and Surgery14(5), 3501–3518 (2024)

    2024

  18. [18]

    Clinics in Derma- tology20(3), 240–247 (2002)

    Johr, R.H.: Dermoscopy: alternative melanocytic algorithms—the ABCD rule of dermatoscopy, Menzies scoring method, and 7-point checklist. Clinics in Derma- tology20(3), 240–247 (2002)

    2002

  19. [19]

    Journal of Basic Engineering82(1), 35–45 (03 1960)

    Kalman, R.E.: A new approach to linear filtering and prediction problems. Journal of Basic Engineering82(1), 35–45 (03 1960)

    1960

  20. [20]

    NPJ Digital Medicine5(1), 171 (11 2022)

    Kline, A., Wang, H., Li, Y., et al.: Multimodal machine learning in precision health: a scoping review. NPJ Digital Medicine5(1), 171 (11 2022)

    2022

  21. [21]

    In: Brazilian Conference on Intelligent Systems

    de Lima, L.M., Krohling, R.A.: Exploring advances in transformers and CNN for skin lesion diagnosis on small datasets. In: Brazilian Conference on Intelligent Systems. pp. 282–296. Springer International Publishing (2022)

    2022

  22. [22]

    In: de Freitas, R., Fur- tado, D

    de Lima, L.M., Rocha, M.B., Krohling, R.A.: A novel interpretable approach to deep multimodal data fusion applied to cancer diagnosis. In: de Freitas, R., Fur- tado, D. (eds.) Intelligent Systems. pp. 35–49. Springer Nature Switzerland, Cham (2026)

    2026

  23. [23]

    Intelligent Medicine3(4), 258–266 (2023)

    de Lima, L.M., de Assis, M.C.F.R., Soares, J.P., Grão-Velloso, T.R., de Barros, L.A.P., Camisasca, D.R., Krohling, R.A.: Importance of complementary data to histopathological image analysis of oral leukoplakia and carcinoma using deep neu- ral networks. Intelligent Medicine3(4), 258–266 (2023)

    2023

  24. [24]

    Expert Systems42(10), e70130 (2025)

    de Lima, L.M., Krohling, R.A.: Hyperbolic space feature enhancement and inter- action mechanism for classification on small multimodal medical datasets. Expert Systems42(10), e70130 (2025)

    2025

  25. [25]

    VMamba: Visual State Space Model

    Liu, Y., Tian, Y., Zhao, Y., Yu, H., Xie, L., Wang, Y., Ye, Q., Jiao, J., Liu, Y.: VMamba: Visual State Space Model. arXiv preprint arXiv:2401.10166 (2024)

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  26. [26]

    In: Proceedings of the 31st International Conference on Neural Information Processing Systems

    Lundberg, S.M., Lee, S.I.: A unified approach to interpreting model predictions. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. p. 4768–4777. NIPS’17, Curran Associates Inc., Red Hook, NY, USA (2017)

    2017

  27. [27]

    Revista Brasileira de Cancerologia72(2), e–025587 (fev 2026)

    Martins, L.F.L., Chaves, G.V., Oliveira, J.F.P., Souza, L.B.L.d., Chagas Neto, P., Carvalho, F.N.d., Vasconcelos, G.M.d., Dias, M.B.K., Mello, M.S.d.C.: Perfil epidemiológico da incidência de câncer no brasil e regiões: Estimativas para o triênio 2026-2028. Revista Brasileira de Cancerologia72(2), e–025587 (fev 2026)

    2026

  28. [28]

    Computers in Biology and Medicine116, 103545 (2020)

    Pacheco, A.G., Krohling, R.A.: The impact of patient clinical information on au- tomated skin cancer detection. Computers in Biology and Medicine116, 103545 (2020)

    2020

  29. [29]

    Data in Brief32, 106221 (2020)

    Pacheco, A.G., Lima, G.R., Salomão, A.S., Krohling, B., Biral, I.P., de Angelo, G.G., Alves Jr, F.C., Esgario, J.G., Simora, A.C., Castro, P.B., et al.: PAD-UFES- 20: a skin lesion dataset composed of patient data and clinical images collected from smartphones. Data in Brief32, 106221 (2020)

    2020

  30. [30]

    Journal of Healthcare Informat- ics Research7(1), 59–83 (Mar 2023)

    Rezk, E., Eltorki, M., El-Dakhakhni, W.: Interpretable skin cancer classification based on incremental domain knowledge learning. Journal of Healthcare Informat- ics Research7(1), 59–83 (Mar 2023)

    2023

  31. [31]

    Cancers15(5), 1421 (2023)

    Song, B., Zhang, C., Sunny, S., Kc, D.R., Li, S., Gurushanth, K., Mendonca, P., Mukhia, N., Patrick, S., Gurudath, S., Raghavan, S., Tsusennaro, I., Leivon, S.T., Kolur, T., Shetty, V., Bushan, V., Ramesh, R., Pillai, V., Wilder-Smith, P., Suresh, A., Kuriakose, M.A., Birur, P., Liang, R.: Interpretable and reliable oral cancer An approach with Mamba to m...

    2023

  32. [32]

    Journal of Medical Internet Research23(11), e22934 (2021)

    Takiddin,A.,Schneider,J.,Yang,Y.,Abd-Alrazaq,A.,Househ,M.,etal.:Artificial intelligence for skin cancer detection: scoping review. Journal of Medical Internet Research23(11), e22934 (2021)

    2021

  33. [33]

    International Journal of Oral Science15(1), 1–44 (2023)

    Tan, Y., Wang, Z., Xu, M., Li, B., Huang, Z., Qin, S., Nice, E.C., Tang, J., Huang, C.: Oral squamous cell carcinomas: state of the field and emerging directions. International Journal of Oral Science15(1), 1–44 (2023)

    2023

  34. [34]

    ACM Comput

    Tay, Y., Dehghani, M., Bahri, D., Metzler, D.: Efficient transformers: A survey. ACM Comput. Surv.55(6), 1–28 (Dec 2022)

    2022

  35. [35]

    arXiv preprint arXiv:2408.06291 (2025)

    Thielmann, A.F., Kumar, M., Weisser, C., Reuter, A., Säfken, B., Samiee, S.: Mambular: A sequential model for tabular deep learning. arXiv preprint arXiv:2408.06291 (2025)

    work page arXiv 2025

  36. [36]

    arXiv preprint arXiv:2411.17207 (2024)

    Thielmann,A.F.,Samiee,S.:OntheefficiencyofNLP-Inspiredmethodsfortabular deep learning. arXiv preprint arXiv:2411.17207 (2024)

    work page arXiv 2024

  37. [37]

    World Health Organization, International Agency for Research on Cancer: Global cancer burden: latest estimates and projections (2024)

    2024

  38. [38]

    Frontiers in Oncology12, 893972 (2022)

    Wu, Y., Chen, B., Zeng, A., Pan, D., Wang, R., Zhao, S.: Skin cancer classification with deep learning: a systematic review. Frontiers in Oncology12, 893972 (2022)

    2022

  39. [39]

    Information Fusion115, 102721 (2024)

    Xua, B., Yang, G.: Interpretability research of deep learning: A literature survey. Information Fusion115, 102721 (2024)

    2024

  40. [40]

    High- Confidence Computing4(2), 100211 (2024)

    Yao, Y., Duan, J., Xu, K., Cai, Y., Sun, Z., Zhang, Y.: A survey on large lan- guage model (LLM) security and privacy: The good, the bad, and the ugly. High- Confidence Computing4(2), 100211 (2024)

    2024

  41. [41]

    Bioengi- neering12(10), 1030 (2025)

    Yao, Z., Yan, Y., Liu, Z., Chen, T., Cho, L., Leung, Y.W., Lu, T., Niu, W., Qiu, Z., Wang, Y., Zhu, X., Wong, K.C.: DermaMamba: A dual-branch vision mamba architecture with linear complexity for efficient skin lesion classification. Bioengi- neering12(10), 1030 (2025)

    2025

  42. [42]

    MedMamba: Vision Mamba for Medical Image Classification

    Yue, Y., Li, Z.: MedMamba: Vision mamba for medical image classification. arXiv preprint arXiv:2403.03849 (2024)

    work page Pith review arXiv 2024

  43. [43]

    Intelligent Medicine3(2), 115–128 (2023)

    Zheng, Y., Li, C., Zhou, X., Chen, H., Xu, H., Li, Y., Zhang, H., Li, X., Sun, H., Huang, X., Grzegorzek, M.: Application of transfer learning and ensemble learning in image-level classification for breast histopathology. Intelligent Medicine3(2), 115–128 (2023)

    2023

  44. [44]

    In: IEEE International Conference on Image Processing

    Zhou, L., Luo, Y.: Deep features fusion with mutual attention transformer for skin lesion diagnosis. In: IEEE International Conference on Image Processing. pp. 3797–3801 (2021)

    2021

  45. [45]

    Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model

    Zhu, L., Liao, B., Zhang, Q., Wang, X., Liu, W., Wang, X.: Vision Mamba: Efficient visual representation learning with bidirectional state space model. arXiv preprint arXiv:2401.09417 (2024)

    work page internal anchor Pith review Pith/arXiv arXiv 2024

This paper was first reviewed by grok-4.3 on June 26, 2026.