Microsoft recently launched a "how-old.net" of the yellow site, do not think crooked, this site pornography does not have a half cents relationship, but in the color of the page used a rare soil yellow system. When a user uploads a personal photo, it can directly calculate the photo of gender and age, this simple and easily understandable Chine natural and detonated the brain hole of netizens, if you these two days not in the circle of friends or microblogging to see friends stint share, it can only show that your friends face is too narrow.
So how does Microsoft use a simple photo to speculate on age and gender? Thanks to Microsoft's new project Oxford Service, which is just released at the BUILD2015 conference, it provides free access to intelligent rest APIs such as images, speech recognition, and SDK for Azure users. According to Microsoft's information disclosed at the BUILD2015 conference, Project Oxford has so far provided an image recognition API, a Face Recognition API, speech recognition and processing APIs, and a semantic recognition service called Luis, where "How-old.net" is to use the image recognition API and the face recognition API.
After the site receives a picture, Microsoft's Project Oxford service first detects the position of the human face, then according to the 27 points on the face to draw conclusions, such as the pupil, Canthus, mouth, nose, etc. obviously will change with the age of the site, of course, compared to pit Dad is (also may be more interesting), " How-old.net "Test results are not very accurate, make-up girls often can be older than the actual 10 years old, but netizens do not care, if really accurate," how-old.net "may not be so quickly popular.
URL: http://how-old.net/