{"id":84,"date":"2016-02-18T11:12:59","date_gmt":"2016-02-18T05:42:59","guid":{"rendered":"http:\/\/ec2-54-169-252-119.ap-southeast-1.compute.amazonaws.com\/?p=84"},"modified":"2024-04-16T12:20:08","modified_gmt":"2024-04-16T06:50:08","slug":"what-is-computer-vision-really","status":"publish","type":"post","link":"https:\/\/www.vue.ai\/blog\/tech-talk\/what-is-computer-vision-really\/","title":{"rendered":"What is Computer Vision, Really?"},"content":{"rendered":"Reading Time: <\/span> 6<\/span> minutes<\/span><\/span>

And how did it get in your apps? Computer vision has been around for many decades, but recently, we\u2019re witnessing a major resurgence of interest in how machines \u2018see\u2019, and how computer vision can be used to build products for consumers and businesses.<\/em><\/strong><\/p>\n

It\u2019s the summer of 1966.<\/p>\n

Seymour Papert, a South African mathematician and computer scientist, has just joined the Artificial Intelligence group at MIT. Next year, he will co-invent the programming language LOGO. But for now, he\u2019s doing something he thinks is much simpler along with his colleague Marvin Minsky: assigning the Summer Vision Project to undergrads.<\/p>\n

The aim of the project is to build a system that can analyze a scene and identify objects in it \u2014 \u2018a real landmark in the development of pattern recognition\u2019.<\/em><\/p>\n

So the vast, puzzling field of computer vision that research scientists and tech giants are still trying to decode was first thought to be simple enough for an undergraduate summer project by the very people who pioneered artificial intelligence.<\/p>\n

\"computer<\/p>\n

The challenge of building a broad visual system turned out to be rather trickier than Papert expected. And it\u2019s still proving difficult, despite all our advances in technology.<\/p>\n

Visual AI is a tough nut to crack<\/h3>\n

Computer vision is not just a way to convert pictures to pixels, and it can\u2019t make sense of a picture just from its pixels. It\u2019s the ability of a machine to take a step back and interpret the big picture that those pixels represent.<\/span>And that\u2019s much harder than we think it is.<\/p>\n

For instance, when we see a picture of a model wearing a dress, we automatically identify which part of the body we\u2019re looking at and from which angle. We can figure out lighting conditions. We may even be able to judge the color and texture of the clothes based on shadows, highlights and color temperature.<\/p>\n

\"visual
Every image in a catalog is different \u2014 backgrounds, pose, lighting, positions, and even image quality can all change. A nightmare for a computer. Images from Voonik.com.<\/figcaption><\/figure>\n

Show the same picture to a computer that\u2019s been trained to identify only people and clothing, and it can easily get sidetracked with shadows, or even fail if a reference it\u2019s been trained to identify is missing or hard to separate from other elements \u2014 for instance, if there\u2019s a decorative gold pattern on the wall directly behind the model that\u2019s similar to the gold embroidery on her dress.<\/p>\n

Or it will, of course, spectacularly fall apart if you switch the photo with one of T.Rex battling King Kong, because it has no clue what a dinosaur or a giant gorilla looks like.<\/p>\n

Computer vision is extremely good at doing specifically defined tasks. But human vision is holistic and can interpret a wide range of things that it detects, something we\u2019re yet to crack with AI. Because tasks that seem easy to us, aren\u2019t quite as simple to model for a computer system.<\/p>\n

As one of our data scientists observed:<\/p>\n

\n

#AI<\/a> does a few things that humans can, but way faster and at scale which is awesome. Ability to do anything a human does is not answered yet<\/p>\n

— Hariharan (@harc007) February 11, 2016<\/a><\/p><\/blockquote>\n