Early Detection of Pneumonia Using Deep Learning on Chest Radiographic Images
Jampala Manasa
*
Information Technology Vasireddy Venkatadri Institute of Technology Guntur, Andhra Pradesh, India.
Palanti Harika
Information Technology Vasireddy Venkatadri Institute of Technology Guntur, Andhra Pradesh, India.
Kanugula Venkata Dyuthisri
Information Technology Vasireddy Venkatadri Institute of Technology Guntur, Andhra Pradesh, India.
Nagababu Pachhala
Information Technology Vasireddy Venkatadri Institute of Technology Guntur, Andhra Pradesh, India.
Koya Amrutha
Information Technology Vasireddy Venkatadri Institute of Technology Guntur, Andhra Pradesh, India.
*Author to whom correspondence should be addressed.
Abstract
There are lots of diseases causing health issues for people, and among them, there is pneumonia which results in deaths of the patients. It is important to recognize symptoms since they cannot be prevented. Chest X-ray can be one of the tools for diagnosing the illness. At the same time, there are some issues concerning this method since it is quite time-consuming and subjective. In this paper, an alternative approach will be discussed, which involves diagnosing pneumonia by means of chest X-rays.
Approach Used and Experiments - In general, the technique suggested in this paper relates to techniques for automated diagnosis of pneumonia. The main principle of the proposed methodology is the use of DenseNet121 neural network and transfer learning. Some preprocessing procedures, including resizing, normalization, and augmentation, should be done before the actual work begins. As the source dataset, we have used images of chest X-rays of healthy people and those diagnosed with pneumonia. The efficiency of this technique is about 85%.
Keywords: Pneumonia detection, deep learning, DenseNet121, chest X-ray, transfer learning, medical image classification, artificial intelligence, healthcare diagnostics