Abstract:

A respiratory infection called pneumonia can be brought on by bacteria or viruses. Numerous people are affected, particularly in emerging and underdeveloped nations where there is a high level of pollution, unhygienic living conditions, congestion, and insufficient medical facilities. make breathing challenging. It results from pneumonia. To ensure effective treatment and boost survival chances, pneumonia must be diagnosed as soon as possible. Chest X-ray imaging is the technique most frequently used to diagnose pneumonia. However, it might be difficult and sensitive to subjective variability to examine chest X-rays. Using chest X-ray pictures, we created a computer-aided diagnosis method this work to automatically detect pneumonia. To deal with the dearth of data available, we used the transfer learning technique. To dealwith the lack of data that is available, we used deep transfer learning and created a Convolutional Neural Network (CNN) model using the four transfer learning techniques CovXNet, RNN, and VGG16. Whereas ResNet 50 is employed in the current approaches, which did not achieve the appropriate accuracy but are improving. So, it is suggested to combine the current strategy with additional transfer learning techniques. On a dataset of pneumonia X-rays that was available to the public, the proposed technique was assessed.

Description:

.