sid0312/OCTImageSegmentation

Image Segmentation of optical coherence tomography images using UNets made with Gluon framework on top of the Apache MxNet

Segmentation of optical coherence tomography images with diabetic macular edema - The Gluon Implementation

Built from scratch using Apache MxNet and Gluon

Plotting the training examples and the results

Performance on our validation examples

The UNet Structure

Model summary. Input -> ndarray of size (5,1,284,284)

Dataset

Images for segmentation of optical coherence tomography images with diabetic macular edema. Obtained the dataset from https://www.kaggle.com/paultimothymooney/chiu-2015

I have included the unzipped version of the dataset in this repository

Installing the requirements

pip3 install -r requirements.txt

Clone the repository

git clone https://github.com/sid0312/OCTImageSegmentation
cd OCTImageSegmentation

Model training

python train.py

Results

python results.py

To get the intuition of the training process, go to https://github.com/sid0312/OCTImageSegmentation/blob/master/unets.ipynb

Made with ❤️ by Siddhant Baldota