Stephen Wolfram, writing on the Wolfram blog:
“What is this a picture of?” Humans can usually answer such questions instantly, but in the past it’s always seemed out of reach for computers to do this. For nearly 40 years I’ve been sure computers would eventually get there—but I’ve wondered when.
I’ve built systems that give computers all sorts of intelligence, much of it far beyond the human level. And for a long time we’ve been integrating all that intelligence into the Wolfram Language.Now I’m excited to be able to say that we’ve reached a milestone: there’s finally a function called ImageIdentify built into the Wolfram Language that lets you ask, “What is this a picture of?”—and get an answer.
And today we’re launching the Wolfram Language Image Identification Project on the web to let anyone easily take any picture (drag it from a web page, snap it on your phone, or load it from a file) and see what ImageIdentify thinks it is.
Here’s a link to the ImageIdentify page. Drop some pictures on it, help it learn.
I dropped an image of Jim onto the page and it said Jim was a person (and not a bearded wonder, which is what I expected). I dropped an image of an Apple Watch and ImageIdentify called it an analog watch.
Needs a bit of work, but definitely interesting.
Finally someone thinking about image analysis properly. First one to make a Photos OSX plugin with this will get rich. I want a system to automatically tag my photos based on some AI.
Flickr’s new image identification is very impressive too. It sorted thousands of my photos into different categories such as people, architecture, animals, and much more. Very fast too. Seems way ahead of Photos for Mac.
“I dropped an image of Jim onto the page and it said Jim was a person (and not a bearded wonder, which is what I expected).”
So Wolfram Language answered the question it promised – WHAT is this? In similar fashion to anyone else who might not know who that ‘well-bearded man’ truly is. Seems like the algorithm is currently at the level of a pre-school or kindergarten child, finding its way from simple descriptors to more complex ones.
It will be interesting to see how fast it moves to more complex identification, such as differentiating a head of lettuce from cabbage, or tomato from apple, or even Jim from Dave!
I suspect Wolfram will be one of the means by which we learn to communicate with our first genuine extraterrestrial contacts.
This is not the first product of the sort. Berkley and Université de Montréal have similar projets on going for years now. And Berkley even has the code on Github.
http://caffe.berkeleyvision.org/
http://deeplearning.net/deep-learning-research-groups-and-labs/
probably — Flickr is a specialist tool/company dedicated to photos and nothing else. Apple Photos is a generalist tool. I don’t expect them to be pushing the edge of photo management, but rather to always stay in “pretty good” range for most people.
Hmmm, quite a few images, one clearly of two women baking, were guessed to be vertebrate, strange… Especially since one was a cake.
Almost got it (NSFW) https://twitter.com/GSStamas/status/598930588674883584
ID’d the original Macintosh as a laser printer. Odd that it can’t get an iconic shape like that correct.