Abstract
In response to the urgent need for swift and accurate diagnosis of COVID-19 during the global pandemic, there has been a growing interest in leveraging advanced technology like deep learning for detection purposes. This project introduces a novel approach termed COVID19-ResCapsNet, which integrates deep neural networks with X-ray imaging data to predict the likelihood of COVID-19 onset in patients. The methodology comprises two key steps. Initially, features are extracted using an enhanced Residual feature extraction network, aimed at capturing crucial information from the X-ray images. Subsequently, multiple Capsule Networks are employed to discern between COVID-19 and non-COVID-19 cases. This hybrid model is assessed using two separate datasets of chest X-ray images: Dataset-1 containing 50 images and Dataset-2 containing 20 images. Remarkably, the proposed method achieves high classification accuracy rates of 0.9988 on Dataset-1 and 0.9933 on Dataset-2 for binary classification tasks. To validate its efficacy, the performance of the COVID19-ResCapsNet model is compared against various state-of-the-art pre-trained classification models. Additionally, the study investigates the impact of two crucial hyper parameters: optimizer selection and batch size. The findings underscore the effectiveness of the proposed COVID-19 detection model, suggesting its potential as a complementary tool in clinical settings for expedited and accurate diagnosis. Keywords: Deep learning, Capsule network, Residual network, COVID-19 detection.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have