De-identifying facial images using singular value decomposition

In this paper, a method is proposed that manipulates images in a manner that hinders face recognition by automatic recognition algorithms. The purpose of this method, is to partly degrade image quality, so that humans can identify the person or persons in a scene, while common classification algorithms fail to do so. The approach used to achieve this involves the use of singular value decomposition (SVD). From experiments it can be concluded that, the method reduces the percentage of correct classification rate by over 90%. In addition, the final image is not degraded beyond recognition by humans.