New Research by Johnny Ho and Colleagues Proposes Use of Deep Learning Techniques to Improve Medical X-Ray Technology

The recent advancements in artificial intelligence, especially convolution neural networks, have made the potential of automated medical image analysis nearly comparable to that of health professionals. In medical imaging, these networks can help medical practitioners diagnose diseases like COVID-19 or pneumonia by highlighting the suspicious regions in chest X-ray film. Many contemporary deep learning techniques only focus on COVID-19 classification tasks using X-ray images, while few attempt to make it explainable with a saliency map.  New research by Turner College management professor Johnny Ho and Fashiar Rahman and Bill Tseng of the University of Texas at El Paso, Michael Pokojovy of Old Dominion University, and Peter McCaffery, Eric Walser, Scott Moen and Alex Vo of the University of Texas Medical Branch, proposes a deep learning approach with image enhancement/cropping and data augmentation to enhance COVID-19 classification. Their deep learning approach adds a multi-layer algorithm for improved visualization in X-ray images and to define and calculate a severity assessment index that measures infection severity. In testing, Ho et al.'s trained model achieved an accuracy score of about 96.5% for the three-class X-ray classification task (i.e., COVID-19, pneumonia, and normal [healthy patients]), outperforming many existing techniques in the literature. The study appears in the current issue of Diagnostics, which is international scholarly open access journal on medical diagnostics.