While it doesn’t grab the big headlines, visual impairment is a catastrophic issue in human health. A study by the World Health Organization (WHO) estimates that 285 million people around the world are living with visual impairments, a number that encompasses 39 million blind people and 246 million people with low vision. 1
In the United States alone, an estimated 10 million Americans are blind or visually impaired, including 5.5 million seniors, according to the National Federation of the Blind. This is a problem that will only grow worse as the population ages. Studies suggest that, over the next 30 years, the current number of blind or visually impaired Americans will double. 2
In the face of numbers like these, it’s clear we need to take aggressive steps to make the world more accessible and hospitable to the blind and visually impaired, while working actively to develop new therapies that prevent vision loss. But we can’t stop there. We also need to work to develop therapies that help the visually impaired recover sight.
This brings us to research that is under way at the Commonwealth Scientific and Industrial Research Organisation (CSIRO) in Australia. The CSIRO Computer Vision group, which is part of organization’s Data61 science business unit, has developed the software for a bionic vision solution designed to restore sight for people with profound vision loss.
This life-changing advance is made possible through new computer vision processing that uses large-scale image datasets to optimize and learn more effective processing. That processing is carried out by a high performance computing (HPC) system called Bracewell, named after an Australian astronomer and engineer who worked in the CSIRO Radiophysics Laboratory during World War II, and whose work led to fundamental advances in medical imaging. 3
The Bracewell supercomputer, built on Dell EMC™ PowerEdge™ servers, provides the extreme scale and processing power researchers need to use large data sets to help train their software to recognize and process more images. Drawing on the combined power of 1,634,304 NVIDIA® CUDA® compute cores and 3,192 Intel® Xeon® compute cores, Bracewell has the ability to compress 126 million years’ worth of calculations into a single second. 4
For the CSIRO Computer Vision group, the ultimate goal is to create vision processing systems that presents sight-impaired people with accurate visualizations of the world in a broad variety of scenarios, to help them interact safely and effectively with the environment around them. 5
This same underlying bionic eye technology has applications beyond that of bringing the gift of sight to visually impaired people. In one case, researchers from the CSIRO Data61 team are partnering with a Chinese self-driving technology company, ZongMu Technology, on a quest to equip vehicles with computer vision. This research builds on CISRO’s work to develop bionic vision. The goal is to enable vehicles to see and understand the environment as humans do — and help them rapidly detect and avoid hazards.
“The idea is to develop computer vision using algorithms to estimate the space between objects according to the vehicle’s motion and predict the potential hazards of moving objects,” according to CSIRO. “Unlike laser sensors which rely on a series of points to identify hazards, computer vision offers richer information and a deeper understanding of road scenes through 3D image analysis.” 6
All of this, of course, takes a lot of computational power — the kind of power found in an HPC system like Bracewell.
“The power of this new system is that it allows our researchers to tackle challenging workloads and ultimately enable CSIRO research to solve real-world issues,” notes CSIRO Deputy Chief Information Officer, and Head of Scientific Computing, Angus Macoustra. 7 “The system will nearly double the aggregate computational power available to CSIRO researchers, and will help transform the way we do scientific research and development.”
These are more reasons why HPC matters — solving real-world issues and transforming scientific research and development. Whether the challenge is to restore human sight or to help self-driving vehicles navigate a busy highway, or to enable machine learning for image recognition or to model new material structures — to name some of the research enabled by the Bracewell system — researchers need the power of high performance computing systems.
Ultimately, HPC makes it possible to solve really hard problems that we couldn’t solve without a massive amount of computing power — problems like the development of bionic vision.
For a closer look at CISRO Bracewell system, and how it enables next-generation research, read “CSIRO powers bionic vision research with new Dell EMC PowerEdge Server-based artificial intelligence capability.”
Making a difference with HPC
High performance computing touches virtually every aspect of our lives. HPC is making weather forecasts more accurate, cancer therapies more precise, fraud protection more foolproof and products more efficient. In this series of articles, we explore these and other use cases that capitalize on HPC and its convergence with data analytics to illustrate why HPC matters to all of us.
1 World Health Organization, “Global Data on Visual Impairments 2010,” © 2012.
2 National Federation of the Blind, “Blindness and Low Vision Fact Sheet.”
3 Dell EMC news release, “CSIRO powers bionic vision research with new Dell EMC PowerEdge Server-based artificial intelligence capability,” July 17, 2017.
4 Andrew Underwood, Dell EMC high performance computing leader, “CSIRO powers next-generation research with Dell EMC PowerEdge-based artificial intelligence capability,” July 18, 2017.
5 George Nott, CIO, “CSIRO doubles compute power with $4m HPC: Will use new Dell EMC system to refine bionic vision efforts,” July 18, 2017.
6 CSIROscope, “Our bionic eye tech helps driverless cars ‘see,’” Feb. 7, 2018.
7 Dell EMC news release, “CSIRO powers bionic vision research with new Dell EMC PowerEdge Server-based artificial intelligence capability,” July 17, 2017.