Time-Frequency Imaging for Epileptic Seizure Detection: Feature Fusion with Transformer Model

Öz

Epilepsy is a common neurological disorder that carries serious risks, such as seizures and irreversible brain damage. Accurate and rapid diagnosis of this condition is of great importance. Traditional EEG signal analysis is manual, time-consuming, and prone to human error. The use of artificial intelligence approaches provides a means for more accurate and faster detection. In this study, EEG signals were converted into 2D images using time-frequency transformation methods. Three image datasets were obtained through time-frequency transformations. Each image dataset was then trained using a transformer model, and feature sets were generated by the model. Different feature sets were combined using the feature fusion method, and these combined sets were classified using machine learning method (support vector machines). With the approach proposed in this study, an overall accuracy of 91.20% was achieved.

Anahtar Kelimeler

• Epileptic Seizure
• Time-Frequency Imaging
• Transformer Model
• Feature Processing

Epilepsi Nöbet Tespiti için Zaman-Frekans Görüntüleme: Transformer Model ile Özellik Füzyonu

Öz

Epilepsi, nöbetler ve bu durumun yol açtığı geri dönüşümsüz beyin hasarı gibi ciddi riskler taşıyan yaygın bir nörolojik hastalıktır. Bu hastalığın doğru ve hızlı bir şekilde teşhis edilmesi büyük önem taşır. Geleneksel EEG sinyal analizi, manuel ve zaman alıcı olup insan hatalarına açıktır. Bu sorunu çözmek için yapay zekâ yaklaşımlarının kullanımı, daha hassas ve hızlı tespit imkânı sunmaktadır. Bu çalışmada, EEG sinyalleri zaman-frekans dönüşüm yöntemleri kullanarak 2B görüntülere dönüştürülmüştür. Zaman-frekans dönüşüm yöntemleri ile üç adet görüntü kümesi elde edilmiştir. Ardından her bir görüntü kümesi transformer model ile eğitilmiştir ve model tarafından özellik setleri oluşturulmuştur. Özellik füzyonu yöntemiyle farklı özellik setleri birleştirilmiş ve bu birleşik setler, makine öğrenmesi yöntemiyle (destek vektör makineleri) sınıflandırılmıştır. Bu çalışmada önerilen yaklaşım sayesinde %91.20 genel doğruluk oranı elde edilmiştir.

Anahtar Kelimeler

• Epileptik Nöbet
• Zaman-Frekans Görüntüleme
• Transformatör Modeli
• Özellik İşleme

Kaynakça

1. C. Günbey and G. Turanlı, “Epilepsi ve Pediatrik Epilepsi Sendromları,” Turkish J. Pediatr. Dis., pp. 1–9, Jan. 2022.
2. S. U. Khan, S. U. Jan, and I. Koo, “Robust Epileptic Seizure Detection Using Long Short-Term Memory and Feature Fusion of Compressed Time–Frequency EEG Images,” Sensors, vol. 23, no. 23, p. 9572, Dec. 2023.
3. R. Tutuk and R. Zengin, “Epileptic seizure detection combining power spectral density and high-frequency oscillations,” Int. J. Appl. Math. Electron. Comput., vol. 11, no. 2, pp. 117–127, Jun. 2023.
4. Y. Pan, X. Zhou, F. Dong, J. Wu, Y. Xu, and S. Zheng, “Epileptic Seizure Detection with Hybrid Time-Frequency EEG Input: A Deep Learning Approach,” Comput. Math. Methods Med., vol. 2022, pp. 1–14, Feb. 2022.
5. M. Varlı and H. Yılmaz, “Kombine Derin Öğrenme Tabanlı Epileptik Nöbet Teşhisi,” Eur. J. Sci. Technol., Oct. 2021.
6. S. Mallick and V. Baths, “Novel deep learning framework for detection of epileptic seizures using EEG signals,” Front. Comput. Neurosci., vol. 18, Mar. 2024, doi: 10.3389/fncom.2024.1340251.
7. S. Srinivasan, S. Dayalane, S. kumar Mathivanan, H. Rajadurai, P. Jayagopal, and G. T. Dalu, “Detection and classification of adult epilepsy using hybrid deep learning approach,” Sci. Rep., vol. 13, no. 1, p. 17574, Oct. 2023.
8. G. Yogarajan et al., “EEG-based epileptic seizure detection using binary dragonfly algorithm and deep neural network,” Sci. Rep., vol. 13, no. 1, p. 17710, Oct. 2023.

9. A. Yadav, S. Rathee, Shalu, D. Sheoran, and P. Kumar, “Advanced Epileptic Seizure Detection Using Deep Learning and Bayesian Optimization,” SN Comput. Sci., vol. 5, no. 7, p. 854, Sep. 2024.
10. Y. Wang, N. Huang, T. Li, Y. Yan, and X. Zhang, “Medformer: A Multi-Granularity Patching Transformer for Medical Time-Series Classification,” May 2024, [Online]. Available: http://arxiv.org/abs/2405.19363
11. C. Forigua, M. Escobar, and P. Arbelaez, “SuperFormer: Volumetric Transformer Architectures for MRI SuperResolution,” Jun. 2024.
12. A. Reigns, “Epileptic Seizure EEG Dataset with CWT Images,” Kaggle Web, 2024. https://www.kaggle.com/datasets/alexreigns14/finaldata
13. K. Khairudin, L. Hakim, F. H. Nasution, and R. Kurniawan, “The Application of Complex Morlet Wavelet for Estimating Damping Ratio and Detecting Inter-Area Oscilation Mode on a Real Power System,” in 2021 International Conference on Converging Technology in Electrical and Information Engineering (ICCTEIE), Oct. 2021, pp. 1–4.
14. B. Bilgehan and C. Kavalcıoğlu, “Sürekli glikoz izleme (CGM) sinyalleri ile tip 1 diyabet tedavisi için sürekli dalgacık dönüşüm (CWT) tabanlı filtreleme yöntemi,” Gazi Üniversitesi Mühendislik Mimar. Fakültesi Derg., vol. 35, no. 2, pp. 581–594, Dec. 2019.
15. D. Wei and Y. Zhang, “Fractional Stockwell transform: Theory and applications,” Digit. Signal Process., vol. 115, p. 103090, Aug. 2021.
16. S. Maksimović and S. Gajić, “Stockwell Transform and its Modifications in Signal Processing Courses: Comparison and Features,” in 2023 22nd International Symposium Infoteh-Jahorına (Infoteh), Mar. 2023, pp. 1–6.
17. İ. Paçal and İ. Kunduracıoğlu, “Data-Efficient Vision Transformer Models for Robust Classification of Sugarcane,” J. Soft Comput. Decis. Anal., vol. 2, no. 1, pp. 258–271, Jun. 2024.
18. M. Agar, S. Aydin, M. Cakmak, M. Koc, and M. Togacar, “Detection of Thymoma Disease Using mRMR Feature Selection and Transformer Models,” Diagnostics, vol. 14, no. 19, p. 2169, Sep. 2024.
19. Y. Alaca, E. Basaran, and Y. Celik, “Enhancing Anomaly Detection in Large-Scale Log Data Using Machine Learning: A Comparative Study of SVM and KNN Algorithms with HDFS Dataset,” ADBA Comput. Sci., Jul. 2024.
20. A. Şener, G. Doğan, and B. Ergen, “A novel convolutional neural network model with hybrid attentional atrous convolution module for detecting the areas affected by the flood,” Earth Sci. Informatics, vol. 17, no. 1, pp. 193–209, Feb. 2024.
21. M. Yildirim, E. Cengil, Y. Eroglu, and A. Cinar, “Detection and classification of glioma, meningioma, pituitary tumor, and normal in brain magnetic resonance imaging using deep learning-based hybrid model,” Iran J. Comput. Sci., vol. 6, no. 4, pp. 455–464, Dec. 2023.
22. A. Çalışkan, “Detecting human activity types from 3D posture data using deep learning models,” Biomed. Signal Process. Control, vol. 81, p. 104479, Mar. 2023.
23. A. T. Karadeniz, E. Başaran, and Y. Çelik, “Classification of walnut dataset by selecting CNN features with whale optimization algorithm,” Multimed. Tools Appl., vol. 83, no. 31, pp. 77061–77076, Feb. 2024.
24. M. Hasnain, M. F. Pasha, I. Ghani, M. Imran, M. Y. Alzahrani, and R. Budiarto, “Evaluating Trust Prediction and Confusion Matrix Measures for Web Services Ranking,” IEEE Access, vol. 8, pp. 90847–90861, 2020.
25. I. Pacal, “Improved Vision Transformer with Lion Optimizer for Lung Diseases Detection,” Uluslararası Muhendis. Arastirma ve Gelistirme Derg., May 2024.
26. S. Zhu et al., “Application and visualization study of an intelligence-assisted classification model for common eye diseases using B-mode ultrasound images,” Front. Neurosci., vol. 18, May 2024.