Skip to content

Latest commit

 

History

History
22 lines (16 loc) · 820 Bytes

README.md

File metadata and controls

22 lines (16 loc) · 820 Bytes

Blindness_detection

Different models were trained to predict the severity of diabetic retinopathy based on oculus fundus images.

The data was obtained from Kaggle in terms of APTOS 2019 Blindness Detection competition.

Some of the models can be found here

The best metrics were obtained for pretrained ResNet-50 model:

  • accuracy 0.844
  • f1 score 0.668

The model can be used with our Telegram bot.

Contributors

Model training

  • Ekaterina Shitik (EkaterinShitik)
  • Denis Bredov (angrygeese)
  • Yuri Slepov (IuriiSl)
  • Yulia Volkova (yvolko)

Telegram bot

  • Ustin Zolotikov (glitchheadgit)