AI Case Study

Google's Verily prioritizes patients by diagnosing retinopathy more accurately than ophthalmologists using deep neural networks

Google's Image recognition AI - Imagenet, can process retinal scans faster and more accurately than (trained) human eye. This enables faster and accurate diagnosis and helps address staff shortage and prioritizing based on severity of the condition.

Industry

Healthcare

Healthcare Providers And Services

Project Overview

Google-Verily uses image recognition AI based on Convolutional Neural Networks to provide access to treatment to patients at risk of blindness caused by diabetes by analyzing retinal scans and providing early diagnosis and prioritizing according to risk.

Reported Results

Proof of concept; results not yet available. However, expected to achieve faster throughput and lower error rates.

Technology

Function

R And D

Core Research And Development

Background

There are more than 415M rpeople living with diabetes at risk of blindness and AI can address the shortage of doctors assisting with early diagnosis and starting treatment on time. Over 45% of patients don't see a doctor until there is some loss of vision.

Benefits

Cost - Reduced waiting times

Data