Abnormality detection in nailfold capillary images using deep learning with EfficientNet and cascade transfer learning DOI Creative Commons
Mona Ebadi Jalal, Omar S. Emam, Cristián Castillo-Olea

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Jan. 15, 2025

Nailfold Capillaroscopy (NFC) is a simple, non-invasive diagnostic tool used to detect microvascular changes in nailfold. Chronic pathological associated with wide range of systemic diseases, such as diabetes, cardiovascular disorders, and rheumatological conditions like sclerosis, can manifest observable the terminal capillaries nailfolds. The current gold standard relies on experts performing manual evaluations, which an exhaustive time-intensive, subjective process. In this study, we demonstrate viability deep learning approach automated clinical screening tool. Our dataset consists NFC images from total 225 participants, normal accounting for 6% dataset. This study introduces robust framework utilizing cascade transfer based EfficientNet-B0 differentiate between abnormal cases within images. results that pre-trained ImageNet dataset, followed by domain-specific classes, significantly enhances classifier's performance distinguishing Normal Abnormal classes. proposed model achieved superior performance, accuracy, precision, recall, F1 score, ROC_AUC 1.00, outperforming both models single convolutional neural network, each attained score 0.67 0.83. demonstrates potential facilitate early preventive measures timely interventions aim improve healthcare delivery patients' quality life.

Language: Английский

Abnormality detection in nailfold capillary images using deep learning with EfficientNet and cascade transfer learning DOI Creative Commons
Mona Ebadi Jalal, Omar S. Emam, Cristián Castillo-Olea

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Jan. 15, 2025

Nailfold Capillaroscopy (NFC) is a simple, non-invasive diagnostic tool used to detect microvascular changes in nailfold. Chronic pathological associated with wide range of systemic diseases, such as diabetes, cardiovascular disorders, and rheumatological conditions like sclerosis, can manifest observable the terminal capillaries nailfolds. The current gold standard relies on experts performing manual evaluations, which an exhaustive time-intensive, subjective process. In this study, we demonstrate viability deep learning approach automated clinical screening tool. Our dataset consists NFC images from total 225 participants, normal accounting for 6% dataset. This study introduces robust framework utilizing cascade transfer based EfficientNet-B0 differentiate between abnormal cases within images. results that pre-trained ImageNet dataset, followed by domain-specific classes, significantly enhances classifier's performance distinguishing Normal Abnormal classes. proposed model achieved superior performance, accuracy, precision, recall, F1 score, ROC_AUC 1.00, outperforming both models single convolutional neural network, each attained score 0.67 0.83. demonstrates potential facilitate early preventive measures timely interventions aim improve healthcare delivery patients' quality life.

Language: Английский

Citations

0