This repository contains code for Stanford MRNet Challenge: Classifying Knee MRIs blogpost.

For details about the dataset and the competition, refer to https://stanfordmlgroup.github.io/competitions/mrnet/

1. Clone the repository.

2. Set up the python(3.7) environment. I prefer conda. Given below is how to create a conda environment by the name mrnet

$ conda create -n mrnet python=3.7
$ conda activate mrnet

Make sure you are always inside the mrnet environment while working with this project.

3. Install dependencies. (You might want to select your virtual environment first)

$ pip install -r requirements.txt

4. Download the dataset (~5.7 GB), and put train and valid folders along with all the the .csv files inside images folder at root directory.

  images/
      train/
          axial/
          sagittal/
          coronal/
      val/
          axial/
          sagittal/
          coronal/
      train-abnormal.csv
      train-acl.csv
      train-meniscus.csv
      valid-abnormal.csv
      valid-acl.csv
      valid-meniscus.csv

5. Make a new folder called weights at root directory, and inside the weights folder create three more folders namely acl, abnormal and meniscus.

$ mkdir weights
$ cd weights
$ mkdir abnormal meniscus acl
$ cd ..

6. All the hyperparameters are defined in config.py file. Feel free to play around those. Especially change the task to train on different diseases.

7. Now finally run the training using python train.py. All the logs for tensorboard will be stored in the runs directory at the root of the project.

• Tensorboard Use tensorboard to view the evaluation results. Run the following command.

$ tensorboard --logdir runs/

or if you are using a remote connection

$ tensorboard --logdir runs/ --host= <host_ip_id>

The dataset contains MRIs of different people. Each MRI consists of multiple images. Each MRI has data in 3 perpendicular planes. And each plane as variable number of slices.

Each slice is an 256x256 image

For example:

For MRI 1 we will have 3 planes:

Plane 1- with 35 slices

Plane 2- with 34 slices

Place 3 with 35 slices

Each MRI has to be classisifed against 3 diseases.

Major challenge with while selecting the model structure was the inconsistency in the data. Although the image size remains constant , the number of slices per plane are variable within a single MRI and varies across all MRIs.

So we are proposing a model for each plane. For each model the batch size will be variable and equal to number of slices in the plane of the MRI. So training each model, we will get features for each plane.

We also plan to have 3 separate models for each disease.

We will be using Alexnet pretrained as a feature extractor. When we would have trained the 3 models on the 3 planes, we will use its feature extractor layer as an input to a global model for the final classification.

-- Neelabh Madan

-- Jatin Prakash

This contains the code for Stanford MRNet Challenge. For more information - visit Stanford MRNet Challenge.

Want to become an expert in AI? AI Courses by OpenCV is a great place to start.