Google’s DeepMind is initiating a new research project to aid doctors spot the early signs of sight-threatening diseases.
The team uses machine learning to analyze more than a million eye scans and thereby creating algorithms that can detect early symptoms that humans might miss. The research project is DeepMind’s second collaboration with the UK’s National Health Service (NHS).
Moorfields Eye Hospital in London will partner with Google and DeepMind for a five-year term as they seek to speed up analysis of eye scans from slow and methodical search to rapid computerized search for a plethora of eye symptoms, like diabetic retinopathy and age-related macular degeneration.
DeepMind’s expertise is deep learning, a version of machine learning that uses algorithms to explore data to interpret data points and relationships. The more data the better with deep learning, which uses powerful computational power to analyze the information. Moorfields has stored more than one million eye scans from patients who have come to the hospital, collected over time, with no individual information attached to them (to preserve privacy). The promise of the research by DeepMind’s study of the Moorfields digital eye scans is its high potential for early detection and treatment of both diabetic retinopathy and age-related macular degeneration.
The partnership between the hospital and Google DeepMind aims to quickly and effectively treat the eye conditions by helping doctors in making a correct and timely diagnosis. In addition, the company hopes to help people avoid preventable diseases.
To take Google DeepMind’s help, the company has decided to provide more than 1 million anonymized eye scans to the company for analysis. The eye scans are historical and therefore, none of the patients currently being treated at the hospital will be affected by the study results. The company will use the scans to and train algorithms to identify the early signs of the disease and other information that doctors usually look for in a patient.
