Last Friday, a video of a Pakistani politician talking to reporters was sent out live over the internet. It was several minutes before anyone noticed that a special video effect was drawing cat faces on the people in the video.
Shaukat Ali Yousafzai is in charge of talking to reporters for the Khyber Pakhtunkhwa province (state) in Pakistan. Volunteers for Mr. Yousafzai’s political party recorded the meeting, sending it out live over the social media site Facebook. Putting the video on the internet was a way of sharing the information with people who couldn’t make it to the meeting.
But by accident, one of the volunteers who was recording the press conference turned on a special filter for the video. A filter is a way of adding a special effect to a picture or a video.
Sadly for Mr. Yousafzai, the filter that got turned on was the “cat filter”. This had the effect of putting pink cat ears, a black nose, and whiskers on whichever person was speaking in the video at the moment.
Though the filter was only on for a few minutes, the video and pictures of it have spread all over the internet. The cat filter has gotten the meeting far more attention than it would have gotten in almost any other way.
Mr. Yousafzai and his party are trying not to make a big deal out of the mistake. They say that they have been leading Pakistan in using new technologies like social media. Mr. Yousafzai said, “The filter was turned on by mistake. Let’s not take everything so seriously.”
That makes sense. It’s not the sort of thing someone would do on purrpose.
Did you know…?
Even though image and video filters are common these days, the technology behind them is complicated.
It’s easy for a person to find a face in a picture and draw something on top of it. It’s much more difficult for a computer program, which can’t actually “look” at the picture.
Instead, an app must use math to study the information used to create the picture. This allows it to locate faces in a picture, even though it still can’t “see” them. Computer programs can now identify faces from almost any angle.
Once the program has found a face, it must locate special features of the face, such as eyes, ears, nose, and mouth. Then the program compares the position of these features to the computer’s idea of a “basic face”.
This allows the program to adjust the spacing to match the face in the picture. This helps make sure that the different parts of the filter (like ears and whiskers) get put in the right place.
The final step for the program is to actually draw the effect (like a cat face) on top of the picture or video. Video can be especially challenging, since the filter must change constantly as the person moves.
In the past, a complicated task like this would have required a massive computer. Now it is possible with handheld devices like phones.
This map has not been loaded because of your cookie choices. To view the content, you can accept 'Non-necessary' cookies.