Parkinson Hastalığının Beyin MRI Görüntülerinden Tespitinde CNN ve Transformer Tabanlı Modellerin Karşılaştırılması

Öz

Parkinson hastalığı, özellikle erken evrelerde belirtilerinin belirgin olmaması nedeniyle tanısı zorlaşabilen önemli nörolojik hastalıklardan biridir. Bu nedenle, hastalığın erken ve doğru şekilde tespit edilebilmesi için bilgisayar destekli yöntemlere olan ilgi giderek artmaktadır. Bu çalışmada beyin manyetik rezonans görüntüleri (MRI) kullanılarak Parkinson hastalığının tespiti amacıyla dört farklı derin öğrenme modeli karşılaştırmalı olarak değerlendirilmiştir. Çalışmada Temel CNN, EfficientViT, DeiT ve MobileViT mimarileri aynı veri kümesi ve ortak eğitim ayarları altında değerlendirilmiştir. Modeller, doğruluk, ağırlıklı F1 skoru, duyarlılık, özgüllük, parametre sayısı, eğitim süresi, çıkarım süresi ve FPS performans metrikleri ile değerlendirilmiştir. Sabit test bölmesinde MobileViT modeli doğruluk, ağırlıklı F1, duyarlılık ve özgüllük metriklerinde 1,0000 değerine ulaşırken, EfficientViT modeli 0,992 doğruluk, 3,35 ms/görüntü çıkarım süresi ve 298,21 FPS ile en iyi hız-performans dengesini sağlamıştır. MobileViT modelinin farklı veri bölmelerindeki performansı ayrıca beş katlı tabakalı çapraz doğrulama ile incelenmiş ve temel performans metriklerinde 1,0000 ± 0,0000 sonucu elde edilmiştir. Elde edilen sonuçlar, özellikle MobileViT ve EfficientViT modellerinin MRI tabanlı Parkinson hastalığı sınıflandırmasında umut verici performans sergilediğini göstermektedir. Bununla birlikte, bu modellerin klinik karar destek sistemlerinde değerlendirilebilmesi için daha geniş, hasta bazlı ve çok merkezli veri kümelerinde doğrulanması ve ileri klinik değerlendirmelerden geçirilmesi gerekmektedir.

Anahtar Kelimeler

• Parkinson hastalığı
• beyin MRI görüntüleri
• derin öğrenme
• transfer öğrenme
• transformer tabanlı modeller

Comparison of CNN- and Transformer-Based Models for the Detection of Parkinson’s Disease from Brain MRI Images

Öz

Parkinson’s disease is an important neurological disorder whose diagnosis can be challenging, especially in the early stages, due to the lack of pronounced symptoms. Therefore, interest in computer-assisted methods for early and accurate disease detection has been increasing. In this study, four deep learning models were comparatively evaluated for the detection of Parkinson’s disease using brain magnetic resonance images (MRI). Specifically, the Baseline CNN, EfficientViT, DeiT, and MobileViT architectures were evaluated using the same dataset and common training settings. Models were evaluated using accuracy, weighted F1 score, sensitivity, specificity, number of parameters, training time, inference time, and FPS performance metrics. On the fixed test set, the MobileViT model achieved a value of 1.0000 in accuracy, weighted F1, sensitivity, and specificity metrics, while the EfficientViT model provided the best speed-performance balance with 0.992 accuracy, 3.35 ms/image inference time, and 298.21 FPS. The performance of the MobileViT model across different data splits was also examined using five-fold stratified cross-validation, yielding a result of 1.0000 ± 0.0000 in key performance metrics. The results indicate that MobileViT and EfficientViT show promising performance for MRI-based Parkinson’s disease classification. However, validation on larger, patient-level, multicenter datasets and further clinical evaluation are required before these models can be considered for use in clinical decision-support systems.

Anahtar Kelimeler

• Parkinson’s disease
• brain MRI images
• deep learning
• transfer learning
• transformer-based models

Kaynakça

1. Boulkrinat NEH, Yahiaoui M, Kaci L. Parkinson's disease detection from brain MRI using convolutional neural networks. Procedia Comput Sci 2024; 251: 660-665.
2. Pathapati RN, Anumala VS, Jupudi VSCG, Pasam N. Improving Parkinson's disease diagnosis: a genetic algorithm-guided CNN approach. In: 2024 2nd International Conference on Signal Processing, Communication, Power and Embedded System (SCOPES); 19-21 December 2024; Paralakhemundi Campus, Centurion University of Technology and Management, Odisha, India. pp. 1-5.
3. Huang Y, Chaturvedi K, Nayan AA, Hesamian MH, Braytee A, Prasad M. Early Parkinson’s disease diagnosis through hand-drawn spiral and wave analysis using deep learning techniques. Information 2024; 15(4): 220.
4. Agrawal S, Sahu SP. Image-based Parkinson disease detection using deep transfer learning and optimization algorithm. Int J Inf Technol 2024; 16(2): 871-879.
5. Desai S, Chhinkaniwala H, Shah S, Gajjar P. Enhancing Parkinson's disease diagnosis through deep learning-based classification of 3D MRI images. Procedia Comput Sci 2024; 235: 201-213.
6. Dogan Y. An innovative approach for Parkinson’s disease diagnosis using CNN, NCA, and SVM. Neural Comput Appl 2024; 36(32): 20089-20110.
7. Li K, Ao B, Wu X, Wen Q, Ul Haq E, Yin J. Parkinson’s disease detection and classification using EEG based on deep CNN-LSTM model. Biotechnol Genet Eng Rev 2024; 40(3): 2577-2596.
8. Siddiqua AM, Oni AM, Miah MJ. A transfer learning approach for neurodegenerative disease classification from brain MRI images: distinguishing Alzheimer's, Parkinson's, and control cases. In: 2024 6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT); 02-04 May 2024; Dhaka, Bangladesh. pp. 347-351.

9. Asgar S, Sahil AA, Uddin MN. Vision transformers with augmented dataset using DDPM to detect Parkinson's disease. In: 2023 26th International Conference on Computer and Information Technology (ICCIT); 13-15 December 2023; Cox's Bazar, Bangladesh. pp. 1-5.
10. Khushbu AN, Yang Z, Liu Y, Zhang X. Parkinson's disease recognition using hybrid CNN-transformer model. In: 2024 4th International Conference on Industrial Automation, Robotics and Control Engineering (IARCE); 15-17 November 2024; Chengdu, China. pp. 307-311.
11. Özdemir EY, Özyurt F. ElasticNet-based vision transformers for early detection of Parkinson’s disease. Biomed Signal Process Control 2025; 101: 107198.
12. Shanthappa PM, Bayari M, Abhilash GB, Gokul KV, Ashish PJ. Parkinson’s disease detection using InceptionV3: a deep learning approach. MethodsX 2025; 14: 103333.
13. Joy MAM, Siddiqua A, Islam MN, Chowdhury MFHK. Automated Parkinson’s disease detection from brain MRI images using deep convolutional neural network. In: 2023 26th International Conference on Computer and Information Technology (ICCIT); 13-15 December 2023; Cox's Bazar, Bangladesh. pp. 1-6.
14. Abdullah SM, Abbas T, Bashir MH, Khaja IA, Ahmad M, Soliman NF, El-Shafai W. Deep transfer learning based Parkinson’s disease detection using optimized feature selection. IEEE Access 2023; 11: 3511-3524.
15. Erdaş ÇB, Sümer E. A fully automated approach involving neuroimaging and deep learning for Parkinson’s disease detection and severity prediction. PeerJ Comput Sci 2023; 9: e1485.
16. Sivaprakasam SA, Pandi SS, Varsha S, Vishnu M, Sharath J. Parkinson’s disease detection: VGG-ResNet hybrid approach. In: 2024 10th International Conference on Communication and Signal Processing (ICCSP); 12-14 April 2024; Melmaruvathur, India. pp. 54-59.
17. Acikgoz H, Korkmaz D, Talan T. An automated diagnosis of Parkinson’s disease from MRI scans based on enhanced residual dense network with attention mechanism. J Imaging Inform Med 2025; 38(4): 1935-1949.
18. Li H, Yang Z, Qi W, Yu X, Wu J, Li H. Parkinson's image detection and classification based on deep learning. BMC Med Imaging 2024; 24(1): 187.
19. Liu X, Peng H, Zheng N, Yang Y, Hu H, Yuan Y. EfficientViT: memory efficient vision transformer with cascaded group attention. In: 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR); 17-24 June 2023; Vancouver, BC, Canada. pp. 14420-14430.
20. Touvron H, Cord M, Douze M, Massa F, Sablayrolles A, Jegou H. Training data-efficient image transformers and distillation through attention. In: Proceedings of the 38th International Conference on Machine Learning; 18-24 July 2021. pp. 10347-10357.
21. Ahamed MS, Malek MA, Debnath A, Reya SS. Classification of wheat leaf diseases using the data-efficient image transformers (DeiT). In: 2024 International Conference on Recent Progresses in Science, Engineering and Technology (ICRPSET); 07-08 December 2024; Rajshahi, Bangladesh. pp. 1-6.
22. Jumphoo T, Phapatanaburi K, Pathonsuwan W, Anchuen P, Uthansakul M, Uthansakul P. Exploiting data-efficient image transformer-based transfer learning for valvular heart diseases detection. IEEE Access 2024; 12: 15845-15855.
23. Mehta S, Rastegari M. MobileViT: light-weight, general-purpose, and mobile-friendly vision transformer. arXiv 2021; arXiv:2110.02178.
24. Li Y, Tang J, Li L, Wang X, Ding W, Li X, Yu T, Wu X. MobileViT-based classification of Alzheimer's disease. In: 2023 IEEE 6th International Conference on Pattern Recognition and Artificial Intelligence (PRAI); 18-20 August 2023; Haikou, China. pp. 443-448.
25. Tang F, Nian B, Ding J, Ma W, Quan Q, Dong C, Yang J, Liu W, ve diğerleri. Mobile U-ViT: revisiting large kernel and U-shaped ViT for efficient medical image segmentation. In: Proceedings of the 33rd ACM International Conference on Multimedia; 27-31 October 2025; Dublin, Ireland. pp. 3408-3417.
26. Sheriff I. Parkinson's Brain MRI Dataset. Kaggle 2023. Available from: https://www.kaggle.com/datasets/irfansheriff/parkinsons-brain-mri-dataset