An Improved Hybrid Deep Learning Approach with Explainable AI for COVID-19 Detection

COVID-19 has posed a significant global challenge since its outbreak. Numerous studies have explored machine learning models to detect COVID-19 from CT scan images, with most relying on generic Convolutional Neural Network(CNN) architectures. However, these models often struggle to capture important features in medical images. In response to this limitation, a customized CNN is developed and combined with pre-trained models such as ResNet, DenseNet and MobileNet. This hybrid model enables more detailed feature extraction. To ensure transparency in decision-making, Explainable AI techniques such as Grad-CAM and LIME are applied. The best model has achieved an accuracy of 98% on the publicly available SARS-CoV-2 CT Scan Dataset, demonstrating its potential for reliable COVID-19 detection in clinical settings. This approach not only improves diagnostic accuracy but also provides clearer insights into the model’s predictions, making it a valuable tool in medical image analysis.