Key Strengths

  • Automatic Real-Time Grading of Diabetic Retinopathy based on the Analysis of Color Fundus Images1
  • Cloud-based, stand-alone, or embedded in most ophthalmology equipments
  • Optimize data processing with real-time image quality assessment2
  • Compatible with any color fundus camera available on the market

Main Features

Image Quality

Image Quality determination based upon:

  • Macula and peripheral vessel visibility
  • Global image sharpness
  • Vessel boundary visibility
  • Vessel density

Automatic Grading of the Diabetic Retinopathy

  • Powerful algorithm based on latest development in Deep Learning
  • Heatmap of most significative pixels to display retinal lesions

Performance

Image Interpretability: 96.4%

 

Sensitivity: 99%

 

Specificity: 87%

 

 

(1) Specificity set to 87.0: Sensitivity = 96.8 % for Idx Inc.; Sensitivity = 99.0% for NotifEye (RetinOptic), Number of false negatives divided by 3

 

(2) Accuracy 96,4%; processing time 0,4s/image on a standard CPU

Scroll to Top