DCT based scalable scrambling method with reversible data hiding functionality
KokSheik Wong, Kiyoshi Tanaka · 2010
This paper proposes a scalable scrambling method in the DCT compressed domain while inserting external information into the content to be scrambled. Scalable scrambling effect is achieved by permuting DC components using a window of size that varies according to a parameter, and external information is encoded by imposing a specific relationship on the cardinality of two adjacent coefficient blocks. The zerorun-level pairs and the cardinality of each coefficient block are generated and en-queued to a list. This list is permuted and sequentially de-queued to generate the scrambled image. The ideas are implemented in JPEG compressed domain, and the basic performance of our scalable scrambling and data hiding method are verified with standard test images.