Google Pedestrian Remover Remains Virtual - For Now
With online privacy an ever-growing concern, University of California compsci grad student Arturo Flores has developed a way to take the people out of the Google.
Right now, Google Maps uses face and licence plate blurring technology whenever a map is zoomed down to Street View. This helps to protect the anonymity of passers by who did not throw on their Sunday best in anticipation of starring in daytime version of "Googleral Hospital" when they left the house.
That one's weak, we know. "Days of our Google" just sounded odd, though.
The trouble with blurry faces, argue some critics, is that they don't do enough to remove the actual identity of pedestrians. Using body shape, clothing and location, it is still possible that identities could be deduced and locations of specific individuals found. This is why 1) you should throw away that purple track suit and 2) always wear clean underwear. You do not want these things on Google Maps.
In response to these concerns, Flores has come up with a way to remove single pedestrians from scenes. His computer algorithm does so by using background imagery of photos taken either before or after the person passed by to approximate and re-pixelate what the background would look like.
As you can see, the results are decent...for the most part.
Some scenes have noticeable jagged edges, and the process tends to leave behind things that the pedestrians were carrying or walking - like umbrellas or dogs,and this can make for an odd viewing experience.
As well, Flores' method cannot yet do groups of people, and does not work in a non-urban setting, or in any photo where the background is not predominately two dimensional.
Overall, though, this is a good idea for addressing the sharply increasing privacy concerns in a globalized world. We do fear that once Google becomes the OneWorld government, however, this will turn into a REAL people remover.
All hail our Google overlords.
Please don't remove us. The track suit is gone. For real. At least, the pants are.