Penumonia
Detector
Using CNN
Name - Anshul Jaiswal
Roll. No. - MCA/45017/23
Presented By: Anshul Jaiswal Get Started
Guided By - Prof. Rakesh Singh Slide 1
Slide 2
Pneumonia Detector
Using Deep Learning
on Chest X-ray
Images
Pneumonia is a lung infection that affects the air sacs
at the end of airway, the infection interfere with the
delivery of oxygen from the air sacs into the blood and
the removal of carbon dioxide from the blood. In this,
basically, it hinder the process of oxygen, absorption,
and the blood, which effectively leads to suffocation.
Slide 3
X-Ray Visibility of Lungs
in Normal vs. Positive
Cases
To identify the pneumonia, the major thing is that in x-ray:
The lungs will be visible for a normal person.
Lungs very not visible or either blurry in case of positive.
Slide 4
NORMAL PNEUMONIA POSITIVE CASE
Slide 5
Objectives of
Pneumonia
Detector Using
CNN
The main objectives of the project are as follows:
Automate the process of detecting pneumonia
from chest X-ray images.
Improve diagnostic speed and accuracy by using
deep learning models.
Develop a scalable solution that can be integrated
into healthcare systems.
Provide a user-friendly interface for healthcare
professionals to upload and analyze X-rays.
Slide 6
Convolutional
Neural Network
A Convolutional Neural Network (CNN) is a type of
deep learning model specifically designed to
process and analyze visual data, such as images.
It uses layers like convolutional layers to
automatically detect patterns (e.g., edges,
textures) and max-pooling layers to reduce the
size of the data, making it highly effective for
image recognition, classification, and other tasks
that involve spatial information.
Slide 7
Slide 8
The chest X-ray dataset is downloaded from Kaggle, which has
Data already been pre-filtered and labeled. It contains images
Collection labeled as either pneumonia-positive or healthy.
All images are resized to a uniform dimension (e.g., 224x224
Preparing pixels) to match the input size expected by the neural network.
Data Pixel normalization is applied, scaling pixel values to a range
(typically 0 to 1) to ensure consistent inputs for the network.
A Convolutional Neural Network (CNN) is used for feature
Model extraction from the chest X-ray images. The architecture
Architecture includes multiple convolutional layers, followed by max-pooling
layers, and fully connected layers for classification.
The dataset is split into training, validation, and testing sets. The
Model CNN model is trained on the training set, and validation data is
Training used to fine-tune the model. Testing data is used for final
evaluation.
Slide 9
The implementation of the Pneumonia Detector involved
designing a CNN model, training it on a preprocessed
dataset,
and evaluating its performance. Below are the key steps of
the
implementation process:
Software Environment
• Programming Language: Python 3.11.5
• Development Tools: Jupyter Notebook
• Device: MacBook Air M2 (15-inch)
• Processor: Apple Silicon M2
• RAM: 8 GB
• GPU: Integrated Apple M2 GPU
• Operating System: macOS Sequoia
Data pre processing Slide 11
The dataset was loaded and preprocessed using the TensorFlow ImageDataGenerator.
CNN Model Architecture
The CNN model was built using the TensorFlow Keras Sequential API.
Slide 10
Training the model Slide 12
The model was trained for 25 epochs using the fit method.
model evaluation Slide 13
The model was evaluated using the test dataset to measure its performance against predefined
metrics such as accuracy.
Front-end Slide 14
Landing page using React JS.
Front-end
Choose Image (when not defected)
Slide 15
Slide 16
Front-end
Choose Image (when defected)
Slide 17
Slide 18
Slide 19
References
Dataset Reference - Mooney, P. (2018). Chest X-
Ray Images (Pneumonia) [Data set]. Kaggle.
Retrieved [25 November, 2024], from
https://www.kaggle.com/paultimothymooney/che
st-xray-pneumonia.
Book Reference Goodfellow, I., Bengio, Y., &
Courville, A. (2016). Deep Learning. MIT Press. ISBN:
9780262035613.
Slide 20
Thank You
This Pneumonia Detector project aims to enhance healthcare
diagnosis by automating the detection of pneumonia. By
leveraging deep learning models, it can reduce the time and
expertise required for diagnosis, making it especially useful in
areas with limited healthcare resources. The project has the
potential to be scaled and integrated into healthcare systems
globally.
Presented By: Anshul Jaiswal End of Slide
