Automated Detection and Classification of Pulmonary Lung Diseases Using Falcon Optimized BiLSTM Approach

An automated and accurate detection of pulmonary lung diseases has become increasingly critical due to the rising global incidence of illnesses like lung cancer, pneumonia, and Chronic Obstructive Pulmonary Disease (COPD). Manual analysis of medical images is often time consuming, highlighting the urgent demand for efficient solutions in medical imaging. To address this, the proposed work develops a Deep Learning (DL) based methods for classification of pulmonary lung diseases. The preprocessing steps, including image augmentation, resizing, normalization, and data splitting is essential to ensure the lung images are appropriately prepared for further analysis. K-means clustering based segmentation is employed for isolating the relevant regions of interest for disease classification. For classification, a Bidirectional Long Short-Term Memory (BiLSTM) network is utilized to capture both temporal and spatial dependencies in the lung images. To further enhance the performance of BiLSTM model, it is fine-tuned using Falcon Optimization Algorithm (FOA), improving classification accuracy of (96%) prediction. The validation from the python software outcomes addressing the need for improved diagnostic capabilities in healthcare.