Downloads: 1 | Views: 74 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1
Research Paper | Computer Science and Information Technology | India | Volume 13 Issue 1, January 2024 | Rating: 5.3 / 10
A Conv - MBDNN Model for the Classification and Detection of Pneumonia Using Transfer Learning on Chest X-Ray Images
Parthasarathy V [2] | Saravanan S. [5]
Abstract: Detecting and classifying pneumonia in X - ray images involves leveraging advanced technologies like Computer Vision (CV) and Machine Learning (ML) to identify and categorize respiratory ailments. The amalgamation of these methodologies holds the potential to aid healthcare professionals in swiftly recognizing and managing pneumonia cases, facilitating early diagnosis and effective treatment. The use of Deep Learning (DL) for the classification and detection of pneumonia in X - ray images has emerged as a successful approach for automated disease identification. As a subset of Artificial Intelligence (AI), DL concentrates on training neural networks with intricate layers to autonomously learn complex representations and patterns from data. In this study, we employ the Conv - MBDNN model to develop the Automatic Detection and Classification of Pneumonia disease. The proposed approach comprises multiple stages aimed at enhancing accuracy and diagnostic performance. Initially, a pre - processing stage is implemented, involving image resizing and the application of a Wiener Filter (WF) to enhance image quality. Subsequently, the feature extraction is conducted using the MobileNetV3 architecture. Finally, the extracted features are fed into an SVM classifier model to distinguish between different types and severity levels of pneumonia. A comprehensive analysis of experimental results demonstrates that the Conv - MBDNN technique outperforms other recent approaches in terms of performance and accuracy in pneumonia detection and classification using X - ray images.
Keywords: Deep Learning, MobileNet, Lung Disease, Pneumonia, Machine Learning, Transfer Learning, Computer Vision
Edition: Volume 13 Issue 1, January 2024,
Pages: 1376 - 1380