Authors : Tejasri Kurapati; Siri Valluri; Lalitha Mulakaluri; Sunanda Datla

Volume/Issue : Volume 5 - 2020, Issue 11 - November

Google Scholar : http://bitly.ws/9nMw

Scribd : https://bit.ly/3qmmhcC

Abstract : Pneumonia is the single largest infectious disease claiming the lives of 2.5 million people, including 672,000 children in 2019 alone. The mortality rate can be significantly reduced if early detection followed by proper treatment is made available. Our project aims to make early detection of pneumonia possible in remote and rural places that lack proper access to skilled radiologists. The diagnosis of this disease is predominantly done by studying chest X-rays. We built an application that can detect pneumonia by scanning chest X-ray images on mobile phones. We developed a convolutional neural network to detect pneumonia in chest X-rays. We converted the neural network into a TensorFlow Lite model to integrate it into an edge device application to enable on-device inference. Through this application, we also propose to help governments identify areas with high infection rates by collecting location data points of users.

Keywords : Pneumonia; Convolutional Neural Network; Edge Devices; Chest X-Ray; TensorFlow Lite; Deep Learning Inference.