Student builds system to remove people from Google Street View images

Physorg reports that Arturo Flores, a student at the University of California, San Diego, has built a computer vision system that removes pedestrians from Google Street View images. Mr. Flores' system used existing detection software to locate pedestrians and replaced their pixels with pixels of the corresponding background taken from other images. The technique works best when the backgrounds are planar. The project was done for a computer vision course taught by Professor Serge Belongie and has been published in the IEEE International Workshop on Mobile Vision.