“AN IMPROVED MODEL FOR PNEUMONIA DETECTION USING CONVOCATIONAL NEURAL NETWORK”

Abstract

Pneumonia continues to be a significant global health concern, necessitating precise and timely detection for effective treatment. In this research, we introduce an innovative algorithm for pneumonia detection utilizing Convolutional Neural Networks (CNNs). Our model integrates a Faster R-CNN architecture with Feature Pyramid Network (FPN) and a Conv-Dilated Net framework, harnessing the deep learning capabilities of CNNs for robust feature extraction and hierarchical representation learning. This integrated approach enables accurate localization and classification of pneumonia lesions in chest X-ray images. The algorithm commences with preprocessing steps aimed at enhancing image quality and normalizing features. Subsequently, convolutional, and max-pooling layers, activated by Rectified Linear Unit (ReLU) functions, are applied to extract features. These feature maps are then fed into dense layers, culminating in an output neuron activated by a sigmoid function for binary classification. To address overfitting, data augmentation techniques are employed during model training, ensuring generalizability to unseen data. Experimental results showcase the effectiveness of the proposed algorithm in accurately detecting pneumonia, achieving high sensitivity and specificity rates. The automation of the diagnosis process through deep learning facilitates prompt and reliable assessment, enabling timely intervention and improving patient outcomes. Future research endeavours will concentrate on refining and validating the model across diverse patient demographics and healthcare environments, with the ultimate objective of integrating it as a valuable clinical tool for pneumonia detection and management.