A deep learning approach to direct immunofluorescence pattern recognition in autoimmune bullous diseases
Open accessAbstract
Background: Artificial intelligence (AI) is reshaping healthcare, using machine and deep learning (DL) to enhance disease management. Dermatology has seen improved diagnostics, particularly in skin cancer detection, through the integration of AI. However, the potential of AI in automating immunofluorescence imaging for autoimmune bullous skin diseases (AIBDs) remains untapped. While direct immunofluorescence (DIF) supports diagnosis, its manual interpretation can hinder efficiency. The use of DL to classify DIF patterns automatically, including the intercellular (ICP) and linear pattern (LP), holds promise for improving the diagnosis of AIBDs. Objectives: To develop AI algorithms for automated classification of AIBD DIF patterns, such as ICP and LP, in order to enhance diagnostic accuracy, streamline disease management and improve patient outcomes through DL-driven immunofluorescence interpretation. Methods: We collected immunofluorescence images from skin biopsies of patients suspected of having an AIBD between January 2022 and January 2024. Skin tissue was obtained via a 5-mm punch biopsy, prepared for DIF. Experienced dermatologists classified the images as ICP, LP or negative. To evaluate our DL approach, we divided the images into training (n = 436) and test sets (n = 93). We employed transfer learning with pretrained deep neural networks and conducted fivefold cross-validation to assess model performance. Our dataset’s class imbalance was addressed using weighted loss and data augmentation strategies. The models were trained for 50 epochs using Pytorch, achieving an image size of 224 × 224 pixels for both convolutional neural networks (CNNs) and the Swin Transformer. Results: Our study compared six CNNs and the Swin Transformer for AIBD image classification, with the Swin Transformer achieving the highest average validation accuracy (98.5%). On a separate test set, the best model attained an accuracy of 94.6%, demonstrating 95.3% sensitivity and 97.5% specificity across AIBD classes. Visualization with Grad-CAM (class activation mapping) highlighted the model’s reliance on characteristic patterns for accurate classification. Conclusions: The study highlighted the accuracy of CNNs in identifying DIF features. This approach aids automated analysis and reporting, offering reproducibility, speed, data handling and cost-efficiency. Integrating DL into skin immunofluorescence promises precise diagnostics and streamlined reporting in this branch of dermatology.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The Swin Transformer generative model performed best across direct immunofluorescence pattern classes (intercellular, linear, negative) for autoimmune bullous disorders.
“For autoimmune bullous disorders, Capurro et al (2024) compared models for automatic classification of direct immunofluorescence patterns, finding that the Swin Transformer generative model performed the best across the pattern classes (intercellular, linear, negative).”
From Artificial intelligence–enabled precision medicine for inflammatory skin diseases (2026) · cites this paper as Capurro et al, 2024
The Automatic classification results section reports that the Swin transformer outperformed the six CNNs on overall validation accuracy, and the paper evaluates negative, intercellular and linear patterns. It describes Swin as a vision transformer used for classification, not a generative model; it also does not report that Swin was the best model separately in each pattern class. Page numbers are unavailable in our copy. (LLM, conf=0.89)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · see citing works
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., & Guo, B. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00986 · Cited by 35,534 · see citing works
PyTorch: An Imperative Style, High-Performance Deep Learning Library
Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al. (2019) · arXiv · doi:10.48550/arxiv.1912.01703 · Cited by 31,473 · see citing works
Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization
Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.74 · Cited by 24,513 · see citing works
(2016) · Cited by 11,679 · see citing works
A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects
Li, Z., Liu, F., Yang, W., Peng, S., & Zhou, J. (2022) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2021.3084827 · Cited by 5,273 · see citing works
Medical image analysis using convolutional neural networks: a review
Anwar, S. M., Majid, M., Qayyum, A., Awais, M., Alnowami, M., & Khan, M. K. (2018) · Journal of Medical Systems · doi:10.1007/s10916-018-1088-1 · Cited by 1,431 · see citing works
Deep convolutional neural network based medical image classification for disease diagnosis
Yadav, S. S., & Jadhav, S. M. (2019) · Journal of Big Data · doi:10.1186/s40537-019-0276-2 · Cited by 1,069 · see citing works
Transfer learning for medical image classification: a literature review
Kim, H. E., Cosa‐Linan, A., Santhanam, N., Jannesari, M., Maros, M. E., & Ganslandt, T. (2022) · BMC Medical Imaging · doi:10.1186/s12880-022-00793-7 · Cited by 1,036 · see citing works
Transformers in medical image analysis
He, K., Chen, G., Li, Z., Rekik, I., Yin, Z., Wen, J., Gao, Y., Wang, Q., Zhang, J., & Shen, D. G. (2022) · Intelligent Medicine · doi:10.1016/j.imed.2022.07.002 · Cited by 571 · see citing works
Deep learning approaches to biomedical image segmentation
Haque, I. R. I., & Neubert, J. J. (2020) · Informatics in Medicine Unlocked · doi:10.1016/j.imu.2020.100297 · Cited by 464 · see citing works
Cross-Entropy Loss Functions: Theoretical Analysis and ApplicationsPreprint
Mao, A., Mohri, M., & Zhong, Y. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2304.07288 · Cited by 337 · see citing works
Mall, P. K., Singh, P. K., Srivastav, S., Narayan, V., Paprzycki, M., Jaworska, T., & Ganzha, M. (2023) · Healthcare Analytics · doi:10.1016/j.health.2023.100216 · Cited by 310 · see citing works
Artificial Intelligence in Dermatology: A Primer
Young, A. T., Xiong, M., Pfau, J., Keiser, M. J., & Wei, M. L. (2020) · Journal of Investigative Dermatology · doi:10.1016/j.jid.2020.02.026 · Cited by 249 · see citing works
Transfer learning for medical images analyses: A survey
Yu, X., Wang, J., Hong, Q., Teku, R., Wang, S., & Zhang, Y. (2022) · Neurocomputing · doi:10.1016/j.neucom.2021.08.159 · Cited by 237 · see citing works
Joly, P., Richard, C., Gilbert, D., Courville, P., Chosidow, O., Roujeau, J., Beylot-Barry, M., D'Incan, M., Martel, P., Lauret, P., et al. (2000) · Journal of the American Academy of Dermatology · doi:10.1067/mjd.2000.107488 · Cited by 227 · see citing works
Autoimmune bullous skin diseases, pemphigus and pemphigoid
Egami, S., Yamagami, J., & Amagai, M. (2020) · Journal of Allergy and Clinical Immunology · doi:10.1016/j.jaci.2020.02.013 · Cited by 174 · see citing works
Towards Better Explanations of Class Activation Mapping
Jung, H., & Oh, Y. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00137 · Cited by 112 · see citing works
Laboratory diagnosis of pemphigus: direct immunofluorescence remains the gold standard
Giurdanella, F., Diercks, G. F. H., Jonkman, M. F., & Pas, H. H. (2016) · British Journal of Dermatology · doi:10.1111/bjd.14408 · Cited by 49 · see citing works
Practical Direct Immunofluorescence
Kim, R. H., & Brinster, N. K. (2020) · American Journal of Dermatopathology · doi:10.1097/dad.0000000000001516 · Cited by 31 · see citing works
Hocke, J., Krauth, J., Krause, C. D., Gerlach, S., Warnemünde, N., Affeldt, K., van Beek, N., Schmidt, E., & Voigt, J. (2023) · Frontiers in Immunology · doi:10.3389/fimmu.2023.1111172 · Cited by 14 · see citing works
Pastore, V. P., Touijer, L., Capurro, N., Cozzani, E., Gasparini, G., Parodi, A., & Odone, F. (2023) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi53787.2023.10230494 · Cited by 4 · see citing works
Self-citation
Flagged first, then by citations
Research works printed but not linked: 2
- Huang G, Liu Z, van der Maaten L, Weinberger KQ. Densely Connected Convolutional Networks. In: 2017 [cited 2023 Sep 21]. p. 4700-8. Available from: https://openaccess.thecvf.com/content_cvpr_2017/html/Huang_Densely_Connected_Convoluti onal_CVPR_2017_paper.html
- Curran Associates, Inc.; 2019. p. 8024-35. Available from: http://papers.neurips.cc/paper/9015-pytorch-an-imperative-style-high-performance-deep- learning-library.pdf
Cited by
Sorted by citations
Cao, C., Zhao, W., Guo, J., Wang, Z., Lü, A., & Liang, C. C. (2025) · The Innovation Medicine · doi:10.59717/j.xinn-med.2025.100154 · Cited by 14 · see citing works
The Diagnostic Classification of the Pathological Image Using Computer Vision
Matsuzaka, Y., & Yashiro, R. (2025) · Algorithms · doi:10.3390/a18020096 · Cited by 12 · see citing works
Exploring the challenge and value of deep learning in automated skin disease diagnosis
Liu, R., Chen, Z., Yao, G., & ZHANG, P. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.109589 · Cited by 8 · see citing works
Artificial intelligence–enabled precision medicine for inflammatory skin diseases
Tang, A. S., Wei, M. L., Haemel, A., La, C., Sirota, M., & Lee, E. Y. (2026) · Journal of Investigative Dermatology · doi:10.1016/j.jid.2025.10.596 · Cited by 6 · see citing works
Exploring AI as a Diagnostic Tool in Medical Imaging for Dermatopathological Diseases
Kakada, P., Ramani, P., Rajinikanth, V., & Janaki, V. R. (2026) · Indian Journal of Dermatology · doi:10.4103/ijd.ijd_341_25 · Cited by 2 · see citing works
AI-Enhanced Diagnosis for Immunological Disorders
Neguț, I., & Vișan, A. I. (2025) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3693-9725-1.ch003 · Cited by 2 · see citing works
Bursting the bubble on diagnostics: artificial intelligence in autoimmune bullous disease
Doolan, B. J., & Thomas, B. (2024) · British Journal of Dermatology · doi:10.1093/bjd/ljae197 · Cited by 2 · see citing works
Kharisma, I. L., Munandar, R. A., Maulana, D. I., Rizky, M., & Kamdan, K. (2025) · International Global Conference Series on ICT Integration in Technical Education & Smart Society · doi:10.3390/engproc2025107052 · Cited by 1 · see citing works
Kakada, P., Yuwanati, M. B., & Ramani, P. (2025) · Journal of Oral Biology and Craniofacial Research · doi:10.1016/j.jobcr.2025.10.023 · Cited by 1 · see citing works
Applications of artificial intelligence in rare skin diseases and skin cancers: A scoping review
Alkhateeb, M., Nayeem, A., Alokla, M., Hajj, N. E., Alam, T., Ahmed, A., Belhouari, S. B., Sheikh, J. I., & Abd-Alrazaq, A. A. (2026) · Intelligence-Based Medicine · doi:10.1016/j.ibmed.2026.100408 · Cited by 1 · see citing works
Dynamic learning rate adaptation via momentum-guided gradient discrepancy
Xiang, Q., Chai, W., Lei, L., Song, Y., & Li, C. J. (2026) · Expert Systems with Applications · doi:10.1016/j.eswa.2026.133676 · Cited by 1 · see citing works
Fan, Z., Li, Z., Jin, Y., Chandrasekaran, A. P., Shakir, I. M., Zhang, Y., Siddique, A., Wang, M., Zhou, X., Tian, Y., et al. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.12.05.627041
Artificial intelligence-enabled precision medicine for inflammatory skin diseasesPreprint
Tang, A., Wei, M. L., Haemel, A. K., La, C., Sirota, M., & Lee, E. Y. (2025) · PubMed · doi:10.48550/arxiv.2505.09527 · Cited by 5 · see citing works
Skin Cancer Detection via Deep Learning Techniques
Naeemah, M. R., & Kamil, M. Y. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-032-03740-4_41
Rasheed, A., Shirazi, S. H., & Alaklabi, S. (2026) · IEEE Access · doi:10.1109/access.2026.3661729
Evaluation of Transformer and Convolutional Neural Networks for Autoimmune Skin Diseases
Roumeissa, L., Sara, S., & Soumeya, Z. (2025) · International Conference on Research in Computing at Feminine (RIF) · doi:10.1109/rif68108.2025.11406835
Fine-grained and multi-pattern anti-nuclear antibody recognition: A new dataset and framework
Ma, C., Yan, Y., Ma, Z., Liang, K., Fan, K., Jin, W., Zhang, X., Wang, Z., & Li, Y. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104075
Artificial Intelligence and Bullous Dermatoses
Karampinis, E., & Zafiriou, E. (2026) · Springer eBooks · doi:10.1007/978-3-032-15614-3_13
Gilda, A. J., & Sethukarasi, T. (2026) · Expert Review of Clinical Immunology · doi:10.1080/1744666x.2026.2663042
Autoimmune Disorder Classification Using CNN Model
Yadav, A., Patil, R., Newaskar, D., Kadam, B., & Rohokale, V. (2026) · International Conference on Computing Communication Control and automation · doi:10.1109/iccubea69821.2026.11702113
Wang, H., Zhang, J., Cui, T., He, Y., Zhang, W., & Jin, H. (2026) · Dermatology Practical & Conceptual · doi:10.5826/dpc.160s1a7713
Albayrak, F. K., Gedik, M. G. M., Zerde, B. O. Z. B. O., Güney, C. A., Özhamam, E. U., Kılınç, F., & Aktaş, A. (2026) · Dermatology Practical & Conceptual · doi:10.5826/dpc.160s1a7708
- Authors
- Capurro, N., Pastore, V. P., Touijer, L., Odone, F., Cozzani, E., Gasparini, G., & Parodi, A.
- Affiliations
- University of Genoa (IT)
- Corresponding author
- Capurro, Niccolò
- ORCID
- Venue
- British Journal of Dermatology
- Publisher
- Oxford University Press (OUP)
- Published
- 2024-04-06
- Volume
- 191
- Issue
- 2
- Pages
- 261-266
- ISSN
- 0007-0963 (ISSN-L)
- 1365-2133
- OpenAlex
- W4394014620
- Cited by
- 22
- References
- 22
- Citations by year
- 2026: 12 · 2025: 8 · 2024: 2
- FWCI
- 6.69
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- British Journal of Dermatology (publishedVersion, cc-by-nc)
- PubMed (publishedVersion)
- CINECA IRIS Institutial Research Information System (University of Genoa) (submittedVersion)
- Topics
- Health Sciences › Medicine › Pathology and Forensic Medicine › Autoimmune Bullous Skin Diseases
- Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Keywords
- autoimmune bullous skin diseases, direct immunofluorescence, linear patterns, Swin Transformer, convolutional neural networks, automated image analysis, dermatological diagnosis, Grad-CAM
- MeSH terms
- Deep Learning, Algorithms, Autoimmune Diseases, Biopsy, Humans, Skin, Skin Diseases, Vesiculobullous, Fluorescent Antibody Technique, Direct
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
