A pattern recognition system has been developed by two researchers in India who have announced that it can tell the difference between an orange and a lemon with 99% accuracy. Shiv Ram Dubey and Anand Singh Jalal will be publishing the proof in this month’s International Journal of Applied Pattern Recognition.
In other news, human supermarket shoppers were relieved to learn that they’re still slightly more intelligent than your average computer. Just when you think computers already know everything, you find out that they’ve been bluffing all along and that they couldn’t even tell the difference between your garden variety citrus fruit.
But, hey, the problem of pattern recognition in computers has been a difficult one. For instance, when three researchers from Google tried to train their computers to recognize images in videos, it took a network of 16,000 processors working together in a neural network of one billion connections to…wait for it…learn to identify a cat. I bet you and I watched way fewer You Tube videos before we figured it out.
Heck, I bet a few internet-savvy rats would catch on faster.
And then there’s the problem of facial recognition. Researchers and security experts often make bold claims about the ability of computers to recognize the human face. Those programs do such a fantastic job that my slender, ten-years-younger Chinese-American friend was identified by my name, and the security team never did admit that they were wrong. (I’m Caucasian and busty, if you must know.)
Part of the reason that some people are opposed to computerized drones fighting our wars is because of fears that they might just hit the wrong person. Oops. They seem to be doing better in recent years, but pattern recognition technology will still take all the successes it can get.
Being able to tell an orange from a lemon doesn’t sound like much. But the researchers assure us that the systems are working up to bigger things.
Soon they’ll have a pattern recognition program that can tell whether or not the orange has a soft spot on it.