Nicholas Bedworth, founder of SocialEyes, shares how they are developing cost-effective mobile AI diagnostic tools that can provide critically-needed medical services to low-resource societies.
“Being able to diagnose people near to where they live – at home, at work, in the community, pharmacies – we can catch these diseases very early before they develop into extremely difficult to treat, sight threating, life threatening conditions within a matter of five, ten, fifteen years,” mentioned Bedworth on the importance of intervening early to prevent chronic diseases.
Their MARVIN (Mobile Autonomous Retinal Evaluation) deep learning-based service is trained with GeForce GTX 1080 GPUs on over 100,000 retinal images to identify and classify chronic diseases, such as diabetic retinopathy, glaucoma and macular edema. The trained neural networks are then deployed using TensorRT on Jetson TX1 to develop their mobile medical device that can diagnose the consequences of the chronic diseases in primary care or community care.
“The main trick is to replicate the medical skills of specialists — tertiary care physicians such as Ophthalmologists — who are in very scarce supply worldwide,” said Bedworth whose team at SocialEyes comprises over 40 Ph.D. and M.S. level domain experts.
Share your GPU-accelerated science with us at http://nvda.ws/2cpa2d4 and with the world on #ShareYourScience.
Watch more scientists and researchers share how accelerated computing is benefiting their work at http://nvda.ws/2dbscA7