neuradetect_py has four AI models (Breast Density Classifier, BIRADS Classifier, Breast Cancer Classifier, GMIC) which will give predictions and visualization by using tensorflow and PyTorch.

This file covers all requirments and dependencies from scratch for application neuradetect_py

follow the below steps to run models

Step 1

Create pubsub topic and subscription

Go to Google Cloud Platform
Click on Topics, create a new topic
create a subscription for the topic
Click on Subscriptions, select topic and create subscription
In Google Cloud Platform select storage click on browser and create bucket to store artifacts eg. credential file etc
create another bucket to store logs.

Note

• Create two topic to run all models parallelly first topic will process 3 scripts and publish message for second topic to process another 2 scripts create three subscriptions of first topic and two subscriptions of second topic
  • Topic-1
    • Topic-1-subscription-1
    • Topic-1-subscription-2
    • Topic-1-subscription-3
  • Topic-2
    • Topic-2-subscription-1
    • Topic-2-subscription-2

Create a vm instance and run commands on SSH terminal window

Recommendation is to have N1 8vcpu configuration to process all models smoothly

Step 2

Follow steps in sequence from bootstrap.sh & install all dependencies

Application flow diagram

How backend of NeuraDetect work ?

• From original model we modified the model script, totally there are five scripts density_model_torch_modified.py, birads_prediction_torch_modified.py, crop_mammogram_extract_centers.py, run_model.py, run_model_gmic.py from four different repository's. All scripts have to be run separately.
• when user will upload images from front end then message with stoarge information of images will get generated and sent to pubsub topic-1.
• Subscriptions of topic-1 one will run scripts (density_model_torch_modified.py, birads_prediction_torch_modified.py, crop_mammogram_extract_centers.py) accordingly and crop_mammogram_extract_centers.py will pass message information in pubsub tpoic-2 and subscription of that topic-2 will process remaining two script( run_model.py, run_model_gmic.py).
  last two scripts are dependend on (crop_mammogram_extract_centers.py) script.

Models

Model Name	Repository
Breast Density Classifier	Model
BIRADS Classifier	Model
Breast Cancer Classifier	Model
GMIC	Model