Recently, computer scientist Stephen Wolfram has released a new tool called the Wolfram Image Identification Project. With this tool, users first need to upload or link to a picture, and then they can see how the computer recognizes objects in the picture.

Wolfram describes his tools like this: Unlike many computer vision projects, the Wolfram project is built around an "artificial neural network." Artificial neural networks are actually a software framework inspired by the biological brain that enables image recognition. The "artificial neural network" constructed by Wolfram can be "trained" by tens of millions of tag images.

"We don't need to use old methods to describe an object, such as a chair. All we do is to provide many types of chairs. You can also say, 'anything we want to define as a chair'." Will put the 'chair' label close to the 'chair' we defined, and the other will be excluded," Wolfram said.

The Business Insider website has found through testing that in most cases, the image recognizer is able to capture at least the "key points" in the image. For example, it can accurately classify Hillary Clinton and Marisa Mayer as "people," but it does not recognize them by name. It is able to classify Paris cafe Le Comptoir as “a building”.

However, in some cases, the Wolfram recognition tool completely ignores the people in the picture and focuses on certain inanimate objects. For example, in a boxing match picture by Gennady Golovkin, the algorithm is locked in boxing gloves and provides additional information about the pair of boxing gloves, such as weight and material.

In another KFC advertising image, the algorithm ignores the hero inside, but focuses on the car next to the hero and identifies it as "traffic."

For Samsung phones, the tool recognizes it as "remote control," and like the boxing gloves mentioned above, the system also provides some textual introductions.

In this Tesla car picture, the system noticed the person standing in front of the car (Tesla CEO Musk), but it classified the car behind Musk as a "two-door" sedan instead of a four-door car. .

In another video game Doom4 screenshot, the system identified the soldier in the picture as a "spider" and provided information about the name and maximum age of the spider.

Torsion Spring

application: viation,automotive, motorbike, bicycle, automatic equipment, household appliance, medical device ,toy, telecommunication,computer, instrument and meter,furniture, electric tools,mold and other industries

China Torsion Spring,Garage Spring manufacturers, welcome Helical Torsion Spring,Double Torsion Spring purchasers from worldwide to visit our site.

Torsion Spring,Garage Spring,Helical Torsion Spring,Double Torsion Spring,Torsion Spring Manufacturer

Shenzhen Lanejoy Technology Co.,LTD , http://www.grill-mesh.com