Innovative Hybrid CNN+SVM Model for Accurate Covid-19 Detection From CT Images

The advent of advanced deep learning techniques has revolutionized various fields, including healthcare, where accurate and efficient diagnostic tools are of paramount importance. In the context of the COVID-19 pandemic, the need for rapid and precise diagnosis is critical to managing and mitigating the spread of the virus. In this study, we propose a decision support system for the diagnosis of COVID-19 using CT images, employing deep learning algorithms. To evaluate the performance of our models, we create a unique dataset that is meticulously curated and tailored to the task at hand. This dataset consists of a large number of CT images categorized into COVID-19 positive and negative classes, allowing for a robust evaluation of our models' capabilities. Our approach involves the development of novel CNN models as well as the exploration of pre-trained architectures, such as ResNet50v2 and VGG16, in a comprehensive modelling study. Additionally, we introduce a hybrid model by combining CNN models with the SVM algorithm. Hyperparameter optimization is performed using the grid search method, and the modelling process utilizes an original dataset with two classes (COVID-19 and Normal). Performance evaluation involves dividing the dataset into training and test sets (85%-15% ratio) and employing 5-fold cross-validation. Proposed novel CNN models achieve an accuracy rate of 99.93% and 99.86%, while the hybrid CNN+SVM model achieves an accuracy rate of 100% and 99.77%, respectively. Successful application of these proposed deep learning models in healthcare shows their potential to improve diagnostic accuracy and patient outcomes.