blackbird71SR / Brain-Segmentation-and-Tumor-Detection

Modified VGG16 and UNetCNN based 4D Image Segmentation (Finalist - Smart India Hackathon 2019)

image-segmentation unet deep-learning convolutional-neural-networks python vgg16 brats-dataset brats-challenge gui django mrbrains18 image-classification medical-imaging

Brain Image Segmentation

Smart India Hackathon -2019 Finalist

Worked on Following Medical Datasets:

BraTS 2018
MrBrains18

Models:

Modified VGG16
Modified UNetCNN

Interface:

GUI using Python PyQt with the ability to visualize the results along with accuracy in real-time
The website made using Django that works the same as GUI but with easy and better visualization

Results:

Segmented parts of the brain in the 4D dataset with average dice score of 0.84.
Able to find the tumor in the dataset with average dice score of 0.92

Examples:

Tumor Detection

Brain Segmentation

Interface

About

Modified VGG16 and UNetCNN based 4D Image Segmentation (Finalist - Smart India Hackathon 2019)

image-segmentation unet deep-learning convolutional-neural-networks python vgg16 brats-dataset brats-challenge gui django mrbrains18 image-classification medical-imaging

Languages

Language:Python 95.5%Language:HTML 1.8%Language:CSS 1.2%Language:Shell 1.1%Language:PowerShell 0.3%Language:JavaScript 0.1%Language:Roff 0.0%