Artificial Intelligence for Disease Detection using Transformers

Abstract

هى يشض ح فُس شذ ذَ انعذوي سَببه COVID-19 هَعب الاكخشاف ان بًكش و SARS-COV- ف شُوط 2 دوسا يه اً ف انس طُشة COVID- انذق قُ نحالاث 91 عه ا خَشاس انف شُوط وحقذ ىَ انشعا تَ انطب تُ ان اًُسبت .ف انس ىُاث الاخ شُة ,اظهشث حق اُُث انخعهى انع قًُ خَائج واعذة ف ححه مُ انصىس انطب تُ وحص فُُ ان هًاو انخ حساعذ ف ححذ ذَ وحشخصُ الايشاض وان شًاكم انصح تُ (CNN) بذقت كب شُة. انشبكاث انعصب تُ انخلاف فُ تُ ىًَرجا ي انخعهى انع قًُ ( ViTs ) ويحىلاث انشؤ تَ انخ حى حطب قُاها ب جُاح عه حص فُُ انصىس انطب تُ. انهذف ي هزا انع مً اجشاء دساست شايهت ويقاس تَ ب , COVID- ف حص فُُ 19 VIT و CNN دقت اًَرج COVID- كذساست حانت نهخ ضًُُ ب حالاث 19 وايشاض انجهاص انخ فُس الاخشي .اجش جَ هز انذساست ححخى عه COVID- عه قىاعذ انب اُ اَث عايت ل 19 صىس الاشعت ان قًطع تُ COVID-19 is a highly contagious respiratory disease caused by the SARS-CoV-2 virus. Early and accurate detection of COVID-19 cases plays a crucial role in controlling the spread of the virus and providing appropriate medical care. In recent years, deep learning techniques have shown promising results in medical image analysis and classification tasks, helping to identify and diagnose diseases and health problems with great accuracy. Convolutional neural networks (CNNs) and visual transformers (ViTs) are two popular deep learning architectures that have been successfully applied to medical image classification. The aim of this work is to carry out a comparative study between the accuracy of ViT and CNN models in the classification of COVID-19, as a case study to distinguish between cases of COVID-19 and other respiratory diseases. This study is carried out on a public COVID-19 database containing CT-Scan images.