Abstract
Mobile Health (mHealth) uses mobile devices and wireless technology to support healthcare practices, enabling widespread access to health services. Recent advancements in machine learning (ML) have enhanced healthcare by improving disease diagnosis and monitoring. However, integrating and executing machine learning models, especially those based on image processing and deep learning, on mobile devices can be challenging due to limited processing power and storage capacity. This study describes the steps of developing an ML-based model and its integration into mobile devices and cloud environments. A skin disease predictor using the MobileNet architecture was developed as a use-case mHealth application. Techniques such as transfer learning, data augmentation, and focal loss were employed to enhance model performance. The mHealth model was then integrated into a mobile device and the cloud environment. The on-device model exhibited faster prediction times (average 108.3 ms) compared to the cloud-based model (average 1281.2 ms). While on-device deployment ensured data privacy and offline functionality, the cloud approach provided scalability and easier updates, but at the expense of latency and data security. By providing a comparative analysis, this work demonstrates the feasibility of integrating ML models into mHealth applications, emphasizing the importance of balancing performance, cost, and usability.
Ethical Statement
The study has not been submitted to another journal for consideration.
Supporting Institution
This research is partially supported by the TUBITAK-BIDEB 2232 International Fellowship program under the grant number 121C085.
Project Number
121C085
Thanks
This research is largely based on the work conducted by Özge Çiçek in her master’s thesis titled "Mobil Sağlık Uygulamalarında Makine Öğrenmesi Temelli Model Geliştirme ve Modelin Cihaz-Bulut Dağıtımı" (Cicek, 2024)
References
- Candemir, S., Nguyen, X. V., Folio, L. R., & Prevedello, L. M. (2021). Training strategies for radiology deep learning models in data-limited scenarios. Radiology: Artificial Intelligence, 3(6), e210014.
- Çiçek, Ö. (2024). Mobil sağlık uygulamalarında makine öğrenmesi temelli model geliştirme ve modelin cihaz-bulut Dağıtımı, Master’s Thesis, Eskisehir Technical University, Türkiye.
- Dai, X., Spasić, I., Meyer, B., Chapman, S., & Andres, F. (2019). Machine learning on mobile: An on-device inference app for skin cancer detection. In IEEE 4th Int. Conf. on Fog and Mobile Edge Computing, pp. 301 305.
- Esteva, A., Kuprel, B., Novoa, R. A., Ko, J., Swetter, S. M., Blau, H. M., & Thrun, S. (2017). Dermatologist-level classification of skin cancer with deep neural networks. Nature, 542(7639), 115-118.
- Goceri, E. (2021). Diagnosis of skin diseases in the era of deep learning and mobile technology., 134, 104458. Computers in Biology and Medicine,134, 104458.
- Han, S., Mao, H., Dally, W.J. (2015). Deep compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding . arXiv preprint arXiv:1510.00149.
- Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., ... & Adam, H. (2017). Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861.
- Jung EY, Kim J, Chung KY, Park DK. (2014). Mobile healthcare application with EMR interoperability for diabetes patients. Cluster Comput 17(3):871–880.
- Khan, Z., Alotaibi, S. (2020). Applications of artificial intelligence and big data analytics in m-health: A healthcare system perspective. Journal Of Healthcare Engineering, 8894694. https://doi.org/10.1155/2020/8894694.
- Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017). ImageNet classification with deep convolutional neural networks. Communications of the ACM, 60(6), 84-90.
- Lin, T. Y., Goyal, P., Girshick, R., He, K., Dollar, P. (2017). Focal Loss for Dense Object Detection. In Proceedings of the IEEE International Conference on Computer Vision (pp. 2980-2988)
- Molina-Recio, G., García-Hernández, L., Castilla-Melero, A., Palomo-Romero, J. M., Molina-Luque, R., Sánchez-Muñoz, A. A., & Salas-Morera, L. (2015). Impact of health apps in health and computer science publications. A systematic review from 2010 to 2014. In Bioinformatics and Biomedical Engineering, Granada, Spain, April 15-17, 2015. Proceedings, Part II 3 (pp. 24-34).
- Sama, P. R., Eapen, Z. J., Weinfurt, K. P., Shah, B. R., & Schulman, K. A. (2014). An evaluation of mobile health application tools. JMIR mHealth and uHealth, 2(2), e3088.
- Silva, B. M., Rodrigues, J. J., de la Torre Díez, I., López-Coronado, M., & Saleem, K. (2015). Mobile-health: A review of current state in 2015. Journal of biomedical informatics, 56, 265-272.
- Susanto, A., Winarto, H., Fahira, A., Abdurrohman, H., Muharram, A., Widitha, U., Efirianti, G., George, Y. Tjoa, K. (2022). Building an artificial intelligence-powered medical image recognition smartphone application: What medical practitioners need to know. Informatics In Medicine Unlocked, 32:101017.
- Tschandl, P., Rosendahl, C. & Kittler, H. (2018). The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Sci. Data, 5, 180161 doi:10.1038/sdata.2018.161.
- Weiss, K., Khoshgoftaar, T. M., & Wang, D. (2016). A survey of transfer learning. Journal of Big data, 3, 1-40.
MOBİL SAĞLIK UYGULAMALARINDA MAKİNE ÖĞRENMESİ TEMELLİ MODEL GELİŞTİRME VE MODELİN CİHAZ-BULUT ENTEGRASYONU
Abstract
Mobil Sağlık (mHealth), mobil cihazlar ve kablosuz teknolojiyi kullanarak sağlık hizmetlerini destekleyen uygulamaları içerir ve sağlık hizmetlerine yaygın erişim sağlar. Makine öğrenimi (ML) alanındaki son gelişmeler, hastalık teşhisi ve takibini iyileştirerek sağlık hizmetlerini geliştirmiştir. Ancak, özellikle görüntü işleme ve derin öğrenmeye dayalı makine öğrenimi modellerini mobil cihazlara entegre etmek ve çalıştırmak, sınırlı işlem gücü ve depolama kapasitesi nedeniyle zor olabilir. Bu çalışma, bir ML tabanlı modelin geliştirilmesi ve mobil cihazlar ile bulut ortamlarına entegrasyonu adımlarını açıklamaktadır. Örnek mHealth uygulaması olarak MobileNet mimarisi kullanılarak bir cilt hastalığı tahmin modeli geliştirilmiştir. Model performansını artırmak için transfer öğrenimi, veri artırma ve focal loss gibi teknikler kullanılmıştır. Eğitilmiş mHealth modeli daha sonra bir mobil cihaza ve bulut ortamına entegre edilmiştir. Cihaz üzerindeki model, bulut tabanlı modele kıyasla daha hızlı tahmin süreleri (cihaz üzerindeki model ortalama 108,3 ms) sergilemiştir (bulut tabanlı model ortalama 1281,2 ms). Cihaz üzerindeki dağıtım, veri gizliliği ve çevrimdışı işlevsellik sağlarken, bulut yaklaşımı ölçeklenebilirlik ve daha kolay güncellemeler sunmuş, ancak gecikme süresi ve veri güvenliği açısından dezavantajlar yaratmıştır. Bu çalışma ML modellerinin mHealth uygulamalarına entegrasyonunun uygulanabilirliğini göstermekte, karşılaştırmalı bir analiz sunmakta ve performans, maliyet ve kullanılabilirlik arasında bir denge kurmanın önemini vurgulamaktadır.
Project Number
121C085
References
- Candemir, S., Nguyen, X. V., Folio, L. R., & Prevedello, L. M. (2021). Training strategies for radiology deep learning models in data-limited scenarios. Radiology: Artificial Intelligence, 3(6), e210014.
- Çiçek, Ö. (2024). Mobil sağlık uygulamalarında makine öğrenmesi temelli model geliştirme ve modelin cihaz-bulut Dağıtımı, Master’s Thesis, Eskisehir Technical University, Türkiye.
- Dai, X., Spasić, I., Meyer, B., Chapman, S., & Andres, F. (2019). Machine learning on mobile: An on-device inference app for skin cancer detection. In IEEE 4th Int. Conf. on Fog and Mobile Edge Computing, pp. 301 305.
- Esteva, A., Kuprel, B., Novoa, R. A., Ko, J., Swetter, S. M., Blau, H. M., & Thrun, S. (2017). Dermatologist-level classification of skin cancer with deep neural networks. Nature, 542(7639), 115-118.
- Goceri, E. (2021). Diagnosis of skin diseases in the era of deep learning and mobile technology., 134, 104458. Computers in Biology and Medicine,134, 104458.
- Han, S., Mao, H., Dally, W.J. (2015). Deep compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding . arXiv preprint arXiv:1510.00149.
- Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., ... & Adam, H. (2017). Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861.
- Jung EY, Kim J, Chung KY, Park DK. (2014). Mobile healthcare application with EMR interoperability for diabetes patients. Cluster Comput 17(3):871–880.
- Khan, Z., Alotaibi, S. (2020). Applications of artificial intelligence and big data analytics in m-health: A healthcare system perspective. Journal Of Healthcare Engineering, 8894694. https://doi.org/10.1155/2020/8894694.
- Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017). ImageNet classification with deep convolutional neural networks. Communications of the ACM, 60(6), 84-90.
- Lin, T. Y., Goyal, P., Girshick, R., He, K., Dollar, P. (2017). Focal Loss for Dense Object Detection. In Proceedings of the IEEE International Conference on Computer Vision (pp. 2980-2988)
- Molina-Recio, G., García-Hernández, L., Castilla-Melero, A., Palomo-Romero, J. M., Molina-Luque, R., Sánchez-Muñoz, A. A., & Salas-Morera, L. (2015). Impact of health apps in health and computer science publications. A systematic review from 2010 to 2014. In Bioinformatics and Biomedical Engineering, Granada, Spain, April 15-17, 2015. Proceedings, Part II 3 (pp. 24-34).
- Sama, P. R., Eapen, Z. J., Weinfurt, K. P., Shah, B. R., & Schulman, K. A. (2014). An evaluation of mobile health application tools. JMIR mHealth and uHealth, 2(2), e3088.
- Silva, B. M., Rodrigues, J. J., de la Torre Díez, I., López-Coronado, M., & Saleem, K. (2015). Mobile-health: A review of current state in 2015. Journal of biomedical informatics, 56, 265-272.
- Susanto, A., Winarto, H., Fahira, A., Abdurrohman, H., Muharram, A., Widitha, U., Efirianti, G., George, Y. Tjoa, K. (2022). Building an artificial intelligence-powered medical image recognition smartphone application: What medical practitioners need to know. Informatics In Medicine Unlocked, 32:101017.
- Tschandl, P., Rosendahl, C. & Kittler, H. (2018). The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Sci. Data, 5, 180161 doi:10.1038/sdata.2018.161.
- Weiss, K., Khoshgoftaar, T. M., & Wang, D. (2016). A survey of transfer learning. Journal of Big data, 3, 1-40.
Details
| Primary Language | English |
|---|---|
| Subjects | Machine Learning (Other) |
| Journal Section | Computer Engineering |
| Authors | |
| Project Number | 121C085 |
| Publication Date | June 3, 2025 |
| Submission Date | January 21, 2025 |
| Acceptance Date | March 18, 2025 |
| Published in Issue | Year 2025Volume: 28 Issue: 2 |
