Brain Tumor Segmentationa and Visualization

Visualizer

The visualizer was designed using HTML and CSS. The Papaya visualizer is open sourced from

Segmentation

A 3D segmentation model is used to obtain the segmentation of tumor tissues from normal tissues

Tumor Localization

A YOLOv5 model is used to localize the areas of the tumor regions

Tumor Classification

A 3D Classification model is used to classify the tumor regions into the respective classes

Running the models

Clone the repository using

git clone https://github.com/andy2507/brain-tumor-segmentation-and-visualization.git

Loading the models

You will have to update the root_directory and ckpt_path parameters in the files classification.py,segresnet.py and yolo.py to run the app successfully. The models can be found in https://drive.google.com/drive/folders/1yAjUVtIYICDGq4j9hh4IuUEkkKFgyRK1?usp=sharing. Download the files and map the directories accordingly. The root_directory will be the directory the folder is present in

Running the models

First enter the repository folder and install the dependencies using pip install requirements.txt To run, while you are still inside the repository folder, type python app.py into the command line.

Go to the link provided and load a folder containing all four modalities: flair, T1, T1ce T2 using the option in File. Run the models necessary and view the results on the visualizer! Two examples along with the results are present in the repository as well!