Four Harrisburg University professors took home the “Outstanding Achievement Award” as they presented their co-authored research paper, “Diagnosing Heart Disease Types From Chest X-Rays Using A Deep Learning Approach,” at the Conference on Computational Science & Computational Intelligence 2019 last weekend.
Professors Siamak Aram, Roozbeh Sadeghian, Stanley Nwoji, and Iheb Abdellatif’s paper will be published by the IEEE Computer Society. Professors Aram and Nwoji presented the paper at the society’s conference, which took place in Las Vegas Dec. 5-7.
The following is an abstract from the work:“A chest x-ray is the most commonly performed diagnostic examination of heart disease. The interpretation of chest radiographs is crucial to detect a variety of conditions that affect millions all around the world every year. These diseases include cancer, heart and lung diseases, tuberculosis, fibrosis, and others. Trained radiologists perform these difficult time-consuming and challenging interpretation tasks. Misdiagnosis can occur. Computer-aided techniques could lead to more accurate and accessible diagnoses. The present paper used the VGG16 architecture to develop a deep convolutional neural network, which addressed the issue of ensuring accurate diagnoses of various diseases in chest radiography. The model performed well for chest radiography, resulting in high levels of diagnostic accuracy and sensitivity and the ability to classify 14 different diseases.”