DEEP LEARNING-BASED ENSEMBLE APPROACH FOR LUNG DISEASE CLASSIFICATION USING CHEST X-RAY IMAGES

Abstract

Recent technologies development makes it possible to apply artificial intelligence-based techniques in nearly all areas of life. The exactness of deep learning models enables the medical industry to identify and classify a wide a broad range of illnesses. Chest X-rays are prescribed to be safe for diagnosis in a number of circumstances because to the high contagiousness of COVID-19. Among the leading causes of disease and morality worldwide is lung disease, which includes a respiratory infection, TB, and Chronic Obstructive Pulmonary Disease (COPD). For better patient outcomes and efficient treatment, an early and precise diagnosis is essential. However, radiologists and physicians frequently use manual interpretation for classic diagnostic techniques like X-rays, CT scans, and laboratory testing, which can be laborious and prone to human mistake. Deep learning techniques have proven more and more successful in automating and enhancing the reliability of medical diagnostics. This research proposed an ensemble model by incorporating the capabilities of convolutional neural networks and gated recurrent unit to enhance performance for lung disease classification using chest X-rays images. The performance of proposed model is assess by employing the lung disease dataset and compared with other deep learning models such as, Artificial Neural Network (ANN), Long Short-Term Memory (LSTM), Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN). By using the CNN features extraction capabilities and Gated Recurrent Unit (GRU) sequential learning efficiency, the ensemble model outperformed in term of accuracy (0.9721), precision (0.9731), recall (0.9721) and F1score (0.9719). These findings shows the effectiveness of proposed model to improving the accuracy of lung diseases identification.