Comparative Analysis of Deep Learning Methods for Stroke Classification Using ResNet Architectures
Uložené v:
| Názov: | Comparative Analysis of Deep Learning Methods for Stroke Classification Using ResNet Architectures |
|---|---|
| Autori: | Yelken, E., Ceylan, M. |
| Informácie o vydavateľovi: | Springer Science and Business Media Deutschland GmbH |
| Rok vydania: | 2025 |
| Predmety: | Deep Learning, Computed Tomography, Medical Image Analysis, ResNet Models, Stroke Classification |
| Popis: | This study presents a comprehensive evaluation of deep learning methods for classifying strokes using computed tomography (CT) images. Focusing on the comparative analysis of three ResNet architectures—ResNet-50, ResNet-101, and ResNet-152—the models were trained on a specialized dataset provided by the Ministry of Health of the Republic of Turkey. This dataset comprises meticulously annotated brain CT scans, validated by a team of experienced radiologists to ensure precision and reliability in identifying stroke regions. The performance of each model was assessed using several key metrics, including accuracy, precision, recall, and F1 score, to provide a holistic evaluation. Among the models, ResNet-152 demonstrated the highest efficacy, achieving an accuracy of 97.24% and an F1 score of 96.67%. ResNet-50 and ResNet-101 also exhibited commendable performance, with accuracies of 96.34% and 95.63%, respectively, and F1 scores of 91.84% and 88.54%. These results underscore the capability of deeper architectures to capture the complex patterns associated with stroke classification effectively. The findings highlight the potential of deep learning models, particularly ResNet-152, in enhancing diagnostic accuracy and accelerating image analysis processes. Such advancements are critical for developing automated clinical decision support systems, which may significantly improve stroke diagnosis and treatment in healthcare environments. This study not only emphasizes the superiority of deep learning-based approaches over traditional diagnostic methods but also provides a foundation for future work aimed at optimizing and deploying these models in clinical settings. © 2025 Elsevier B.V., All rights reserved. |
| Druh dokumentu: | conference object |
| Jazyk: | English |
| Relation: | EAI/Springer Innovations in Communication and Computing; Konferans Öğesi - Uluslararası - Kurum Öğretim Elemanı; https://hdl.handle.net/20.500.13091/10987; 163; 176; Q3 |
| DOI: | 10.1007/978-3-031-88999-8_13 |
| Dostupnosť: | https://hdl.handle.net/20.500.13091/10987 https://doi.org/10.1007/978-3-031-88999-8_13 |
| Rights: | none |
| Prístupové číslo: | edsbas.223E0C88 |
| Databáza: | BASE |
Nájsť tento článok vo Web of Science | Abstrakt: | This study presents a comprehensive evaluation of deep learning methods for classifying strokes using computed tomography (CT) images. Focusing on the comparative analysis of three ResNet architectures—ResNet-50, ResNet-101, and ResNet-152—the models were trained on a specialized dataset provided by the Ministry of Health of the Republic of Turkey. This dataset comprises meticulously annotated brain CT scans, validated by a team of experienced radiologists to ensure precision and reliability in identifying stroke regions. The performance of each model was assessed using several key metrics, including accuracy, precision, recall, and F1 score, to provide a holistic evaluation. Among the models, ResNet-152 demonstrated the highest efficacy, achieving an accuracy of 97.24% and an F1 score of 96.67%. ResNet-50 and ResNet-101 also exhibited commendable performance, with accuracies of 96.34% and 95.63%, respectively, and F1 scores of 91.84% and 88.54%. These results underscore the capability of deeper architectures to capture the complex patterns associated with stroke classification effectively. The findings highlight the potential of deep learning models, particularly ResNet-152, in enhancing diagnostic accuracy and accelerating image analysis processes. Such advancements are critical for developing automated clinical decision support systems, which may significantly improve stroke diagnosis and treatment in healthcare environments. This study not only emphasizes the superiority of deep learning-based approaches over traditional diagnostic methods but also provides a foundation for future work aimed at optimizing and deploying these models in clinical settings. © 2025 Elsevier B.V., All rights reserved. |
|---|---|
| DOI: | 10.1007/978-3-031-88999-8_13 |