Computers could soon be able to see

Dec 3, 2009 13:28 GMT  ·  By

Endowing computers with artificial vision is something that robotics experts have been after for a long time. However, this is a very complex task that has left many research teams puzzled over how to defeat the numerous challenges ahead. Now, researchers at the Harvard University and the Massachusetts Institute of Technology (MIT) say that they managed to exceed the limitations, by combining molecular biology with advances in high-performance gaming hardware. The team says that genetic screening techniques were also key in providing a way to build better artificial visual systems.

Humans have the innate ability to recognize objects they see, or at least infer some basic characteristics of those they've never met before. But the biological and mental processes that underlie this ability are still a mystery. Researchers admit that they've only begun to probe the depths of the human brain and of the visual cortex. Since they don't know how the human eye works, they find it terribly difficult to mimic this trait inside artificial systems (robotic eyes). A breakthrough in this stall came once experts started using Graphics Processing Units (GPU) for their investigations.

These are the same processors used by the gaming industry to accelerate new computer games, which have tremendous requirements of video cards. The investigators recently published their advancements, which were based on using GPUs as well, in the November 26 issue of the open-access scientific journal PLoS Computational Biology. The work was conducted by Harvard Visual Neuroscience Group expert and principal investigator David Cox, who collaborated closely with MIT PhD candidate Nicolas Pinto, who works in the McGovern Institute for Brain Research and the Department of Brain and Cognitive Sciences at the Institute.

“Reverse-engineering a biological visual system – a system with hundreds of millions of processing units – and building an artificial system that works the same way is a daunting task. It is not enough to simply assemble together a huge amount of computing power. We have to figure out how to put all the parts together so that they can do what our brains can do,” Cox says. “While studying the brain has yielded critical information about how the brain is wired, we currently don't have enough information to build a computer system that works like the brain does. Even if we take all of the clues that we have available from experimental neuroscience, there is still an enormous range of possible models for us to explore,” Pinto adds.

“GPUs are a real game-changer for scientific computing. We made a powerful parallel computing system from cheap, readily available off-the-shelf components, delivering over hundred-fold speed-ups relative to conventional methods. With this expanded computational power, we can discover new vision models that traditional methods miss,” Pinto concludes.