harshini1331 / decoding-pulmonary-health

DECODING PULMONARY HEALTH THROUGH DEEP LEARNING TECHNIQUES

Accurate diagnosis of pulmonary conditions, including critical illnesses such as COVID-19, Bacterial Pneumonia, and Viral Pneumonia, is paramount for effective patient care. This project proposes a comprehensive methodology that harnesses advancements in deep learning techniques to facilitate the precise identification of various pulmonary diseases using chest X-ray images.

About

Pulmonary diseases pose significant challenges to public health worldwide, with conditions such as COVID-19, Bacterial Pneumonia, and Viral Pneumonia causing immense morbidity and mortality. Accurate and timely diagnosis is paramount for effective management and treatment of these conditions. However, traditional diagnostic methods often rely on subjective interpretation and can be time-consuming, leading to delays in patient care and treatment initiation. In recent years, advancements in medical imaging and deep learning have opened up new avenues for improving the accuracy and efficiency of pulmonary disease diagnosis.

Features

• Implements advance neural network method.
• A framework based application for deployment purpose.
• High scalability.
• Less time complexity.

Requirements

• Operating System: Requires a 64-bit OS (Windows 10 or Ubuntu) for compatibility with deep learning frameworks.
• Development Environment: Python 3.6 or later is necessary for coding the sign language detection system.
• Deep Learning Frameworks: TensorFlow for model training.
• Version Control: Implementation of Git for collaborative development and effective code management.
• IDE: Use of Jupyter as the Integrated Development Environment for coding, debugging, and version control integration.
• Additional Dependencies: Includes scikit-learn, TensorFlow (versions 2.4.1), TensorFlow GPU for deep learning tasks.

System Architecture

Output

Output1 - Dataset

Output2 - Distribution of Pixel Intensity

Output3 - Accuracy and Loss Graph

Output4 - Model Comparison

Results and Impact

In this project, we aimed to develop and evaluate deep learning models for the classification of chest X-ray images into four categories: COVID-19, normal, bacterial pneumonia, and viral pneumonia. Through a comprehensive methodology that included data collection, preprocessing, model selection, training, and evaluation, we achieved several key outcomes and insights.

Articles published / References

1. Atitallah SB, et al. (2023) Randomly Initialized Convolutional Neural Network for the Recognition of COVID-19 using X-ray Images. arXiv preprint arXiv:2105a.08199, 2021
2. Breve F (2021) COVID-19 Detection on chest x-ray images: a comparison of cnn architectures and ensembles. arXiv preprint arXiv:2111.09972
3. El Asnaoui K, Chawki Y (2020) Using X-ray images and deep learning for automated detection of coronavirus disease. Journal of Biomolecular Structure and Dynamics: p. 1–12.
4. Mohammed MA, et al. (2022) Novel crow swarm optimization algorithm and selection approach for optimal deep learning COVID-19 diagnostic model. Computational intelligence and neuroscience.
5. Chowdhury ME, et al. Can AI help in screening viral and COVID-19 pneumonia? IEEE Access. 2020;8:132665–132676. doi: 10.1109/ACCESS.2020.3010287.