The next time you see someone with a make up style that puts David Bowie to shame, don’t laugh too much. He or she may be cleverly fooling face recognition and detection systems with a crazy or asymmetrical design.
It’s generally not easy to avoid being spotted by face recognition and detection systems because they use a rather solid algorithm to identify faces:
Based on the so-called Viola-Jones method, the algorithm examines the spatial relationships of an object captured in an image and looks for features commonly found in faces. Most faces have a dark region just above the eyes, while the cheek bones and nose bridge will appear lighter. When the algorithm detects enough such attributes, it guesses the object is a face. The method is generally regarded as effective. Errors are in favour of false positives, making it hard for unobstructed faces to escape notice when they aren’t captured at an angle.
While the algorithm is effective, it can be fooled with make up applied to “alter the contrasts the technology looks for”. Adam Harvey, a student at New York University, has discovered that “dark patterns applied around eyes and cheek bones” do this trick quite well by “throwing of the symmetry” and making you look silly.
While Harvey’s approach isn’t foolproof, it shows that face recognition and detection systems aren’t perfect either. That and the fact that Mr Bowie might’ve been on to something. [The Register]