xx-craziness-xx/pneumonia-detection

Pneumonia Diagnosis using X-ray

Introduction

The data is from Kaggle

There are two category in kaggle's data sets : Normal and Pneumonia

The data is split into a set of 3 folders : train, val and test

Images

Data Exploration

• The train folder totally have 5216 jpg files (Normal:1341，PNEUMONIA:3875)

• The val folder totally have 16 jpg files (Normal:8，PNEUMONIA:8)

• The test folder totally have 624 jpg files (Normal:234，PNEUMONIA:390)

Remark! Data sets for Normal & Pneumonia are imbalanced (about 1:3)

Data Augmentation

ImageDataGenerator(
        rotation_range=10,
        width_shift_range=0.2,
        height_shift_range=0.2,
        rescale=1./255,
        shear_range=0.2,
        zoom_range=0.2,
        horizontal_flip=True,
        fill_mode='nearest')

Screenshots

If you are not getting the app just like the below picture, then change the streamlit theme

Normal

Pneumonia

Deployment

Deployed this project on AWS

http://13.233.229.207:8501/

Model Building

Here I used Xception

Xception Architecture

This is a generic code if you want you can use any other model. All you have to do is replace Xception with your desired model name and last_conv_layer_name with that model's last convolution layer

IMAGE_SIZE=[224,224]
base_model=Xception(input_shape=IMAGE_SIZE + [3],weights='imagenet',include_top=False)

last_conv_layer_name = "block14_sepconv2_act"

for layer in base_model.layers[:-8]:
    layer.trainable=False

new_model = base_model.output
new_model = GlobalAveragePooling2D()(new_model)
new_model = Dense(2,activation='softmax')(new_model)

model=Model(base_model.input,new_model)

Model training

• batch size = 64
• optimizer = adam
• loss = categorical_cross_entropy
• epochs = 30
• steps per epoch = 32

Results

Model Accuracy - 93%

Classification Report

ROC

Run Locally

Clone the project

  git clone https://github.com/xx-CRAZINESS-xx/Pneumonia-Detection.git

Go to the project directory

  cd Pneumonia-Detection

Install dependencies

  pip install -r requirements.txt

Start the server

  streamlit run app.py