Improved Deep Learning Model for Detection and Classification of Pneumonia from X-Ray Images

Title

Improved Deep Learning Model for Detection and Classification of Pneumonia from X-Ray Images

Creator

Sarwath, Unnisa

Contributor

Vijayalakshmi, A

Description

Pneumonia is a severe respiratory disease that can lead to inflammation, fluid accumulation in lungs and breathing difficulties, which needs immediate and accurate diagnosis. Chest X-Ray images are a necessary tool to diagnose pneumonia because manual interpretation poses challenges, particularly for radiologists with less expertise. Artificial intelligence (AI), specifically Convolutional Neural Networks (CNNs), has become a significant in the field of pneumonia detection within chest X-Ray images in recent years. This research presents SarNet, a neural network model developed for the identification of pneumonia in chest X-Ray images. The study involved the compilation of dataset containing chest X-Ray images categorized as normal, pneumonia, and COVID-19 pneumonia cases, each accompanied by appropriate annotations. This dataset was employed as the basis for training and assessing SarNet's performance, underscoring its promise in transforming the diagnosis of pneumonia. SarNet proved highly effective, achieving good accuracy, sensitivity, and specificity compared to traditional diagnostic methods. The model's simplicity, with 41 layers, strikes a balance between depth and computational complexity, enhancing efficiency and ensuring accurate pneumonia detection. Furthermore, the study expanded its scope to include COVID-19 pneumonia detection. SarNet achieved an accuracy of 99.15% in binary classification and 94.9% in multiclass classification, including healthy, pneumonia, and COVID-19 pneumonia cases.

Source

Author's Submission

Date

2024-01-01

Publisher

Christ(Deemed to be University)

Subject

Computer Science

Rights

Open Access

Relation

61000327

Format

PDF

Language

English

Type

PhD

Identifier

http://hdl.handle.net/10603/570001