Title Defense

Progress presentation

Pneumonia detection of chest –ray using

machine learning
Presented by Supervised by

• Syeda Maisha Hasan Jahin • Mondira Chakraborty

Id-211120001 Lecturer
Department of Computer Science
• Tasnin Tabassum Raya & Engineering
Id-211120002 Notre Dame University Bangladesh
 Contents
• Introduction
• Objectives
• Motivation
• Related Works
• Tools
• Implementation
• Future Work

Progress Title Defense

Presentation
17/09/2024 Department of CSE, NDUB 2
 Introduction
● Pneumonia is a lung infection where the air sacs fill with
fluid, causing breathing difficulties, coughing, and fever.
● Pneumonia is an acute lower respiratory disease with traces of
infection usually because of some bacteria or virus.
● 7% of the population or 450 million people worldwide are
suffering from Pneumonia which results in the death of 4
million people every year.[1]

Progress Title Defense

Presentation
17/09/2024 Department of CSE, NDUB 3
 Background study

Progress Title Defense

Presentation
17/09/2024 Department of CSE, NDUB 4
 Motivation
• Early detection of pneumonia can save lives, especially for vulnerable groups like
children and the elderly.
• Traditional diagnosis methods are time-consuming and error-prone, increasing the
need for automated and accurate solutions.
• Automated systems can reduce the burden on healthcare professionals and make
diagnosis more accessible in resource-limited areas.

Title Defense
Progress Presentation
17/09/2024 Department of CSE, NDUB 5
 Objective
• Develop an accurate machine learning model to detect pneumonia from chest X-ray
images.
• Compare the performance of different machine learning algorithms in identifying
pneumonia from X-ray images.
• Evaluate the model's accuracy against traditional diagnostic methods used by
radiologists.
• Explore how the model can help doctors in diagnosing pneumonia earlier and more
accurately.

ProgressTitle Defense
Presentation
17/09/2024 Department of CSE, NDUB 6
 Tools

Machine learning library

Chest -ray dataset

Data processing

Programming language

ProgressTitle Defense
Presentation
17/09/2024 Department of CSE, NDUB 7
 Related Work
Title: Pneumonia Detection Proposing a Hybrid Deep Convolutional Neural Network
Based on Two Parallel Visual Geometry Group Architectures and Machine Learning
Classifiers.
• Technology: CAD, DL, Hybrid CNN, CXR
• Dataset: CXR, Kaggle
• Algorithm: FC layers, SVM, LR, Ensemble

ProgressTitle Defense
Presentation
17/09/2024 Department of CSE, NDUB 8
 Related Work
Title: Pneumonia Detection Using Convolutional Neural Networks.
• Technology: DL, CNN, Transfer learning.
• Dataset: CXR, ImageNet.
• Algorithm: Pre-trained, SVM, RF, LR.

ProgressTitle Defense
Presentation
17/09/2024 Department of CSE, NDUB 9
 Related Work
Title: Pneumonia detection in chest X-ray images using convolutional neural
networks and transfer learning
• Technology:
- CNNs for pneumonia detection using chest X-ray images.
- Transfer learning: VGG16, VGG19, ResNet50, Inception-v3.
• Dataset: Chest X-ray images (binary classification: pneumonia vs non-pneumonia).
• Algorithm: Custom CNN: 2 and 3 convolutional layers.

ProgressTitle Defense
Presentation
17/09/2024 Department of CSE, NDUB 10
 Related Work
Title: Pneumonia Classification using Deep Learning in Healthcare
• Technology:
- CNNs for pneumonia detection using chest X-ray images.
- Transfer learning: VGG16, VGG19, ResNet50, Inception-v3.
• Algorithm: Custom CNN: 2 and 3 convolutional layers.
• Dataset: Chest X-ray images (binary classification: pneumonia vs non-pneumonia).

ProgressTitle Defense
Presentation
17/09/2024 Department of CSE, NDUB 11
 Related Work
Title: Diagnosis of Pneumonia Using Deep Learning
• Technology: AI, Deep Learning, Neural Networks.
• Algorithm: Neural networks for pneumonia detection.
• Dataset: Chest X-ray images.

ProgressTitle Defense
Presentation
17/09/2024 Department of CSE, NDUB 12
 Implementation

ProgressTitle Defense
Presentation
17/09/2024 Department of CSE, NDUB 13
 Future Work
• Enhance model performance by exploring advanced deep learning architectures like ResNet,
DenseNet, or EfficientNet.
• Test the model on larger, diverse datasets to improve its robustness and generalization across different
populations.
• Implement a real-time system for automatic pneumonia detection in clinical settings for rapid
diagnosis and decision support.

ProgressTitle Defense
Presentation
17/09/2024 Department of CSE, NDUB 14
 References
1) Chan H.P, Sahiner B, Hadjiyski L, Zhou C and Petrick N 2005 Lung
nodule detection and classification
2) https://ourworldindata.org/pneumonia (image:slide-5)
3) https://drive.google.com/file/d/1kzdalKKw3FfSoDKk1fjHfVNEAlDmLw
07/view?usp=drivesdk
4) https://drive.google.com/file/d/1dZyHVVjMKl24epgruVX45D5-YmNkF
-0m/view?usp=drivesdk
5) https://drive.google.com/file/d/1MxniWwLieVYHPeabjtEAggNsLk-2bO
Me/view?usp=drivesdk

Progress Title Defense

Presentation
17/09/2024 Department of CSE, NDUB 15
 ANY QUESTIONS

Title Defense
Progress Presentation
17/09/2024 Department of CSE, NDUB 16
 Progress Title Defense
Presentation
17/09/2024 Department of CSE, NDUB 17