Skip to content

Code for Kaggle APTOS 2019 Blindness Detection Challenge

Notifications You must be signed in to change notification settings

submagr/blindness-detection

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

APTOS 2019 Blindness Detection:

Task was to predict the severity of an input image given 3K labeled retina images. We used various preprocessing steps (CLAHE, smoothing, blood vessels segmentation) to remove noise and variable lighting conditions using OpenCV. We experimented with various architectures (ResNet, EfficientNet) and training parameters to improve score using PyTorch. EfficientNet-B2 and TTA gave a final optimized kappa score of 0.895 and final leaderboard ranking of 1309.

Setup

  • Virtual Environment Setup and
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
  • Install Data from kaggle in data\

  • Make virtual environment compatible with Jupyter Notebook

ipython kernel install --user --name=blindness-detection

About

Code for Kaggle APTOS 2019 Blindness Detection Challenge

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 62.0%
  • Jupyter Notebook 38.0%