Detecção de Pneumonia Causada por COVID-19 Utilizando Few-Shot Learning
Abstract
Radiological data are important for the diagnostic and treatment of COVID-19 and other lower respiratory tract infections, such as pneumonia. In this context, we propose in this paper a method that employs a Siamese Network designed to perform Few-Shot Learning to detect pneumonia related to COVID-19 in X-ray images. The generalizability of this model is evaluated using two datasets from different sources, allowing internal and external evaluation. The data partitioning is based on patient identifiers. The model was able to achieve over 96% accuracy, precision, and sensitivity in the internal evaluation. However, in the external evaluation the result was unexpected. On the other hand, it was observed that the Siamese Network model with Few-Shot Learning outperforms a traditional CNN.
References
Cascella, M., Rajnik, M., Aleem, A., Dulebohn, S., and Di Napoli, R. (2022). Features, evaluation, and treatment of coronavirus (covid-19). StatPearls.
Chen, Y., Wang, X., Liu, Z., Xu, H., and Darrell, T. (2020). A new meta-baseline for few-shot learning. CoRR, abs/2003.04390.
Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., and Adam, H. (2017). Mobilenets: Efficient convolutional neural networks for mobile vision applications. CoRR, abs/1704.04861.
Jadon, S. (2021). COVID-19 detection from scarce chest x-ray image data using few-shot deep learning approach. In Park, B. J. and Deserno, T. M., editors, Medical Imaging 2021: Imaging Informatics for Healthcare, Research, and Applications. SPIE.
Li, M. D., Arun, N. T., Aggarwal, M., Gupta, S., Singh, P., Little, B. P., Mendoza, D. P., Corradi, G. C., Takahashi, M. S., Ferraciolli, S. F., Succi, M. D., Lang, M., Bizzo, B. C., Dayan, I., Kitamura, F. C., and Kalpathy-Cramer, J. (2020). Improvement and multi-population generalizability of a deep learning-based chest radiograph severity score for COVID-19.
Roberts, M., , Driggs, D., Thorpe, M., Gilbey, J., Yeung, M., Ursprung, S., Aviles-Rivero, A. I., Etmann, C., McCague, C., Beer, L., Weir-McCall, J. R., Teng, Z., Gkrania-Klotsas, E., Rudd, J. H. F., Sala, E., and Schönlieb, C.-B. (2021). Common pitfalls and recommendations for using machine learning to detect and prognosticate for COVID-19 using chest radiographs and CT scans. Nature Machine Intelligence, 3(3):199-217.
Shorfuzzaman, M. and Hossain, M. S. (2021). Metacovid: A siamese neural network framework with contrastive loss for n-shot diagnosis of covid-19 patients. Pattern Recognition, 113:107700.
Wang, Y. and Yao, Q. (2019). Few-shot learning: A survey. CoRR, abs/1904.05046.
