شادی کبیری، کارشناسی ارشد مهندسی کامپیوتر -شبکه‌های کامپیوتری، دکتر مهدی اکبری کوپایی

نام و نام خانوادگی:شادی کبیری
عنوان پایان نامه: تشخیص بیماری کووید-19 در تصاویر ریه با استفاده از شبکه عصبی عمیق سیامی و الگوریتم فراابتکاری جهت انتخاب ویژگی
رشته تحصیلی:مهندسی کامیپوتر -شبکه‌های کامیپوتری
مقطع تحصیلی: کارشناسی ارشد ناپیوسته
استاد راهنما: دکتر مهدی اکبری کوپایی
چکیده:

Covid-19 disease, which is a member of the coronavirus family, has become a global pandemic and has had many adverse individual and social effects. The disease has affected various aspects of the individual such as health, safety and well-being of individuals and social aspects such as economic losses, unemployment, insufficient medical resources. The most basic and main method of controlling Covid-19 disease to reduce its adverse effects is early diagnosis of the disease to reduce mortality and control its epidemic. In the field of diagnosis of Covid-19 disease, various methods based on data mining techniques, including machine learning and deep learning methods, have been proposed so far. Research shows that in-depth learning can be successful in detecting Covid-19. In this regard, due to the importance of early diagnosis of Covid-19 disease, in this study, a physician assistant system for diagnosing this disease based on deep learning and machine learning is presented. The method proposed in this research is based on the new deep architecture of Siamese neural networks, which consists of two convolutional subnets. Siamese neural networks have been used in the proposed disease diagnosis model to extract deep features from patients’ lung CT images. In this method, to achieve the optimal subset of features that can lead to reduced calculations and early detection of disease in big data, the meta-heuristic algorithm of the Great Pyramid of Giza based on archeology is used and then to diagnose Covid-19 disease from a combination of three versions of the algorithm The nearest neighbor k classification, including its simple, weighted, and fuzzy versions, is used in combination with the majority voting technique. To evaluate this method, a database of CT images of patients’ lungs was used and the results of various experiments show that the method was able to achieve 98.64% accuracy in diagnosing the disease, which was about 11% more successful in diagnosing the disease compared to previous methods.

کلیدواژه: Covid-19 disease, Siamese neural networks, Giza Great Pyramid Algorithm, Fuzzy classification.