4M: Computer Vision with CNN

Code

Note

content/ro

This directory contains materials for computer vision applications using Convolutional Neural Networks (CNN).

Jupyter Notebook

CNN_&_Retinal_optical_coherence_tomography (1).ipynb: Comprehensive CNN tutorial using retinal OCT images

Dataset

Dataset location (Kermany OCT 2018 / OCT2017)

For this training, the dataset is available on the cluster filesystem at:

$FAST/data/kaggle_kermany2018/

Inside that folder you will find the OCT2017/ directory with the standard split:

$FAST/data/kaggle_kermany2018/OCT2017/
├── train/
├── test/
└── val/

Slides

../../Presentation_CV_shared.pptx

Getting Started

To run the CNN notebook:

  1. Ensure the dataset is available at $FAST/data/kaggle_kermany2018/OCT2017/

  2. (optionally) Install Python dependencies: requirements.txt

  3. Launch Jupyter notebook server

  4. Open the CNN notebook and follow the step-by-step guide

Additional Resources