Classification and Predictions of Lung Diseases from Chest X-rays Using MobileNet V2

Featured Application: The method presented in this paper can be applied medical computer systems for supporting diagnosis.Abstract: Thoracic radiography (chest X-ray) is an inexpensive but effective and widely used imaging procedure. However, a lack of qualified radiologists severely limits the applicability technique. Even current Deep Learning-based approaches often require strong supervision, e.g., annotated bounding boxes, to train such systems, which impossible harvest on large scale. In work, we proposed classification prediction lung pathologies frontal thoracic X-rays using modified model MobileNet V2. We considered transfer learning with metadata leverage. NIH Chest-Xray-14 database, did comparison performance our approach other state-of-the-art methods pathology classification. main was by Area under Receiver Operating Characteristic Curve (AUC) statistics analyzed differences between classifiers. Overall, notice considerable spread achieved result average AUC 0.811 accuracy above 90%. conclude that resampling dataset gives huge improvement performance. intended create capable being trained, devices low computing power because they implemented into smaller IoT devices.