Abstract: Lung disease can be considered as the serious health problem in people’s life. Pneumonia is impulsive life-threatening crisis due to lung collapse. Normally it is detected on chest X-ray. The treatment is dependent on timely review of radiographs. Chest radiography (Chest X-ray) is thus far effective, low- cost and broadly used medical imaging trials. The radiologists primarily did the diagnoses manually over each scan, with no automated assistance. The proposed system would greatly improve the efficiency of radiologists, with their knowledge to analyse the chest Xray images. This method uses deep learning approach to predict the Pneumonia lung disease from chest X-ray images. Input is given as chest X-ray image and after pre-processing, it will be fed into Convolutional Neural Network model for disease prediction. The problem can be cast as binary classification problem, where the output is either Pneumonia or Normal. The training model is developed with chest X-rays dataset which contains 5856 chest X-ray images. Keras and TensorFlow are used as tools for implementation.

Keywords: Pneumonia, Chest X-ray, VGG16, ResNet-50, Inception-V3

| DOI: 10.17148/IARJSET.2023.105117