This study presents a novel deep learning-based method, DTH-EPODL, for COVID-19 diagnosis using chest X-ray images. The method incorporates Gaussian filtering for noise removal, MixNet for feature extraction, and a deep autoencoder (DAE) for classification, with hyperparameters optimized via the eagle perching optimizer (EPO). Integration with IoT enables efficient collection and analysis of patient data. Experimental results on benchmark datasets demonstrate superior diagnostic accuracy compared to existing methods.