Softwares Required

For mac and ubuntu ,python 2.7 and Tensorflow1.12. For windows both combination of python 3.6 and python2.7 along with Tensorflow is required to run the project

#Dependencies For list of dependencies check requirement.txt ,there will be other dependencies as well based on python version you are running.Please try to install them based on dependency error you getting.

Xvision

Chest Xray image analysis using Deep Learning and exploiting Deep Transfer Learning technique for it with Tensorflow.

The maxpool-5 layer of a pretrained VGGNet-16(Deep Convolutional Neural Network) model has been used as the feature extractor here and then further trained on a 2-layer Deep neural network with SGD optimizer and Batch Normalization for classification of Normal vs Nodular Chest Xray Images.

Nodular vs Normal Chest Xray

Some specifications

Property	Values
Pretrained Model	VggNet-16
Optimizer used	stochastic gradient descent(SGD)
Learning rate	0.01
Mini Batch Size	20
Epochs	20
2 Layers	512x512
GPU trained on	Nvidia GEFORCE 920M

Evaluation

Confusion Matrix and Training Error Graph

	Normal	Nodule
Precision	0.7755102	0.55555556
Recall	0.76	0.57692308

Accuracy : 69.3333 %

DataSet

openi.nlm.nih.gov has a large base of Xray,MRI, CT scan images publically available.Specifically Chest Xray Images have been scraped, Normal and Nodule labbeled images are futher extrated for this task.

How to use ?

The above code can be used for Deep Transfer Learning on any Image dataset to train using VggNet as the PreTrained network.

Steps to follow

1. Download Data- the script download images and saves corresponding disease label in json format.

python scraper.py <path/to/folder/to/save/images> For eg python scraper.py "D://Xvision/data" ,this will save the images dataset into data folder of Xvision project.You can run python scraper.py parallely also to generate images set faster by giving the range from 0 to 75 or (0-30,30-50,50-75)or based on your preference to generate images fast.

2. Follow the scraper/process.ipynb notebook for Data processing and generate.You can refer run.py in scraper folder or follow the scraper/process.ipynb and run each step individually and generate the below dataset.All the steps of scraper/process.ipynb are done in run.py.You just need to change the folder destination based on your path. After running run.py you should be able to generated below folders.

• Training images folder - All images for training(Eg - final_train_images_calc_nodule_only(folder name))
• Testing images Folder - All images for testing(Eg - final_test_images_calc_nodule_only(folder name))
• Training image labels file - Pickled file with training labels(Eg - training_labels_calc_nodule_only(file name))
• Testing image labels file - Pickled file with testing labels(Eg - testing_labels_calc_nodule_only(file name))

3. Extract features(CNN Codes) from the maxpool:5 layer of PreTrained CovNet(VggNet) and save them beforehand for faster training of Neural network.You have to download vgg16.tfmodel from the tensorflow and save in DeepLearning sub folder or you can download from this link vgg16.model.

   python train.py <Training images folder> <Testing image folder> <Train images codes folder > <Test images codes folder>

   For eg

   python train.py final_train_images_calc_nodule_only final_test_images_calc_nodule_only train-code test-code

4. The extracted features are now used for training our 2-Layer Neural Network from scratch.The computed models are saved as tensorflow checkpoint after every Epoch.

   python train_model.py <Training images folder> <Train images codes folder> <Training image labels file> <Folder to save models>

   For eg

   python train_model.py final_train_images_calc_nodule_only train-code training_labels_calc_nodule_only train-model

5. Finally the saved models are used for making predictions.Confusion Matrix is used as the Performance Metrics for this classifcation task.

   python test_model.py <Testing images folder> <Testing images codes folder> <Testing image labels file> <saved models>

   For eg

   python test_model.py final_test_images_calc_nodule_only test-code testing_labels_calc_nodule_only train-model

Flask REST Server

To run the flask rest server run python app.py in DeepLearning folder and it will run the flask rest server of HealthCare XRAY MRI Analysis. REST API ENDPOINT - /healthcare/mri/analysis (accepts one or more file and return the mri xray analysis)

Some Predictions

References

1. Learning to Read Chest X-Rays: Recurrent Neural Cascade Model for Automated Image Annotation

2. Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning