COVID-19 Identification and Detection from CT-Images using AI Based Ensemble Model

Abstract

The corona virus infection 2019 (COVID-19) has already reached every corner of the globe, leaving many areas with insufficient access to medical supplies. When contrasted to the RT-PCR test, computed tomography (CT) images are able to provide adequate a diagnosis that is both accurate and quick about COVID-19. In this regard, the focus of this research is on the development of an AI-based prediction classifier for the identification and categorization of COVID-19. Ensembles of DL models will be used in the AIEM-DC method in order to accomplish the method's primary goal of accurate COVID-19 detection and classification. In furthermore, a pretreatment approach that relies on Gaussian filtering (GF) is used in order to get rid of clutter and increase image resolution. In addition, for the purpose of extracting the features, a shark optimization method (SOA) is used, along with an array of deep learning methods. These models include recurrent neural networks (RNN), long short-term memory (LSTM), and gated recurrent unit (GRU). In addition, for the categorization of CT images, an upgraded version of the bat method combined with a multiclass support vector machine (MSVM) architecture is used. The originality of the study is shown by the development of the prediction classifier, which includes optimized parameter tuning of the MSVM model for COVID-19 categorization. The usefulness of the AIEM-DC method was tested using a benchmark CT imaging data set, and the findings indicated the potential generalization ability of the AIEM-DC methodology in comparison to the most current state-of-the-art techniques.

Ähnliche Arbeiten