A Novel Reversible Censorship Algorithm

Image files may contain portions of sensitive or intimate content that should not be visible to everyone. This paper deals with various methods of image processing and coding, to protect privacy through face identification. We proposed an encryption procedure implemented in software that allows face detection and subsequent blurring of selected fragments by DCT image transformation while preserving the details needed for successful reconstruction within the same image. A blind steganographic method was used for information embedding into the source cover image. This protected image is accessible to be viewed for ordinary viewers and a dedicated decoder is required for content restoration. The decoder ensures that only authorized persons are allowed to see the content without censorship. The conclusion of the paper focuses on the quality evaluation of the proposed method based on observed statistical parameters.