TY - GEN
T1 - LSB-steganography framework for stereoscopic images based on BJND
AU - Noriega-Galeana, Luis Roberto
AU - Reyes-Reyes, Rogelio
AU - Ponomaryov, Volodymyr
AU - Cruz-Ramos, Clara
N1 - Publisher Copyright:
© 2017 IEEE.
PY - 2017/11/14
Y1 - 2017/11/14
N2 - Nowadays, stereoscopic images are commonly used for 3D image generation, these are composed by a pair of images viewed independently by each eye, which creates a feeling of immersion and in depth perception. 3D vision technologies have multiple applications in many different areas such as medicine, entertainment, computer vision, etc.; being accessible to a wider group of people. However, there is so few researches that seek to hide information in this type of 3D content. This paper proposes a new method to hide information into stereoscopic images using a LSB-Steganography technique and a Binocular Just Noticeable Difference model (BJND) embedding the maximum payload capacity avoiding visual artifacts and inaccurate 3D generation. Results demonstrate that it is possible to embed up to 1.87 Mb into the stereoscopic images keeping high imperceptibility, obtaining average values for PSNR and SSIM between the original stereoscopic pair and the modified of 31.12dB and 0.9693, respectively. Quantity Bad Pixels (QBP, δd= 0. 5) and SSIM are calculated for evaluating accurate 3D generation, obtaining average values of 41.26% and of 0.7721 between the ground-truth and the disparity map of the modified stereoscopic pair.
AB - Nowadays, stereoscopic images are commonly used for 3D image generation, these are composed by a pair of images viewed independently by each eye, which creates a feeling of immersion and in depth perception. 3D vision technologies have multiple applications in many different areas such as medicine, entertainment, computer vision, etc.; being accessible to a wider group of people. However, there is so few researches that seek to hide information in this type of 3D content. This paper proposes a new method to hide information into stereoscopic images using a LSB-Steganography technique and a Binocular Just Noticeable Difference model (BJND) embedding the maximum payload capacity avoiding visual artifacts and inaccurate 3D generation. Results demonstrate that it is possible to embed up to 1.87 Mb into the stereoscopic images keeping high imperceptibility, obtaining average values for PSNR and SSIM between the original stereoscopic pair and the modified of 31.12dB and 0.9693, respectively. Quantity Bad Pixels (QBP, δd= 0. 5) and SSIM are calculated for evaluating accurate 3D generation, obtaining average values of 41.26% and of 0.7721 between the ground-truth and the disparity map of the modified stereoscopic pair.
KW - 3D Images
KW - BJND
KW - Stereoscopic Images
KW - payload capacity
KW - steganography
UR - http://www.scopus.com/inward/record.url?scp=85040628286&partnerID=8YFLogxK
U2 - 10.1109/ICEEE.2017.8108888
DO - 10.1109/ICEEE.2017.8108888
M3 - Contribución a la conferencia
AN - SCOPUS:85040628286
T3 - 2017 14th International Conference on Electrical Engineering, Computing Science and Automatic Control, CCE 2017
BT - 2017 14th International Conference on Electrical Engineering, Computing Science and Automatic Control, CCE 2017
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 14th International Conference on Electrical Engineering, Computing Science and Automatic Control, CCE 2017
Y2 - 20 September 2017 through 22 September 2017
ER -