Chad Catacchio is a contributor writing on a variety of topics in tech. He has held management positions at a number of tech companies in th Chad Catacchio is a contributor writing on a variety of topics in tech. He has held management positions at a number of tech companies in the US and China. Check out his personal blog to connect with him or follow him on Twitter (if you dare).
A University of California, San Diego computer science grad student has created software that can “erase” people from Google Street View images, at least in an urban setting.
The research was shown at the IEEE International Workshop on Mobile Vision as a proof of concept. While Google blurs out faces of pedestrians, other factors including height, neighborhood the person is in, clothes, etc, can make people in Street View recognizable. This software is designed to go those extra steps further to totally wipe out a person from the images, however.
The system is not perfect, however,and mostly only works in urban environments. It is mostly ghost free, but again, it’s not perfect. TechNewsDaily explains it like this:
“The computer vision system replaces holes in the images with an approximation of the actual background behind each pedestrian. These corresponding background pixels are pulled from the image taken right before or right after the image in question.
In addition, the system struggles to generate background pixels when the pedestrian happens to be walking in the same direction as the vehicle at just the right speed. In these cases, the pedestrian may cover up the same spot in multiple frames, foiling the computer scientists’ pixel-swapping approach to removing pedestrians.”
The UCSD research was done by grad student Arturo Flores, and if it can move beyond a proof-of-concept (and possibly beyond Google?), the new unnamed system could have wide-reaching privacy implications. We’ll certainly keep an eye on this very cool story.
Get the TNW newsletter
Get the most important tech news in your inbox each week.