Lossless Data Hiding in the Spatial Domain for High Quality Images Source: IEICE Trans. Fundamentals, Vol. E90-A, No. 4, April 2007, pp. 771-777. Authors: H. L. Jin, M. Fujiyoshi, and H. Kiya
Outline Introduction Proposed Algorithm Experimental Results Conclusions
Send Introduction Reversibly Image Steganography Sender Receiver Secret data: 1001… Cover image Secret data: 1001… Send Cover image Stego-image
Proposed Algorithm = floor((97+100+100+99+102+100+105+105) /8))=101 g A pixel 3*3 block b. each pixel is k=8 bits Overlapped pixel blocks. b = floor((97+100+100+99+102+100+105+105) /8))=101 g
Proposed Algorithm(cont.) ( 2k - 2 = 254) 1. Calculating s : s = the minimum of sb's for all 3x3 blocks. 2. Deciding embeddable blocks: block b is embeddable if |△b| < s. 3. Embedding: (e.g., secret bit wn = 1 --> gb'= 101-2+1 = 100)
Proposed Algorithm(cont.) 4. Extraction: in this case, wn = (100-101) mod 2 = 1. 5. Restoration of gb : in this case, gb = (100+101-1)/2 = 100.
Results
Results(cont.)
Results(cont.)
Conclusion A lossless data embedding method in the spatial domain has been proposed. Only one parameter (the s) needs to be kept. Improves hidden capacity and stego-image's quality.