基於分碼多重存取之可逆式資料隱藏技術

CDMA-Based Reversible Data Hiding Schemes

黃宇任(Yu-Ren Huang)

指導教授 ： 黃育銘

國立暨南國際大學/科技學院/資訊工程學系/碩士(2021年)

https://doi.org/10.6837/ncnu202100181

摘要

本文探討基於分碼多重存取 (Code Division Multiple Access; CDMA) 技術的可逆式資料隱藏演算法。從沃爾什轉換矩陣中取出的行或列做為展頻碼，利用展頻碼之間互為正交的特性，使同一個向量可重疊地藉由不同的展頻碼嵌入多個秘密位元，假設秘密位元是均勻分佈，在多重嵌入時，多個不同的所嵌入之展頻序列，大部份序列元素間值的加總會是0，所以可達到高嵌入量但對影像品質的破壞不會很劇烈。
本文以[5]為基礎，首先提出基於CDMA的另一種嵌入順序與修改方法，接者對[5]所採用之菱形預測誤差(rhombus-error-prediction)做修改，使其所產生之預測誤差可以更符合CDMA資料嵌入時的需求，最後再加入[9]所提之基於菱形預測誤差的變異數排序繼續地提升所提方法的效能。
實驗結果顯示，對於lena影像，相較於[5]的嵌入順序及方法，在PSNR為48.13 dB 時，本文所提之嵌入順序及方法會增加0.0405 bpp的嵌入量，而再套用了本文所提對修菱形預測誤差做修改，會再有0.0075bpp嵌入量的提升，最後再套入菱形預測誤差的變異數排序後PSNR會有所提升，如在嵌入量同為0.0625bpp時，PSNR會有0.78dB的提升。

關鍵字

分碼多重存取 ； 預測誤差 ； 變異數排序 ； 可逆式資料隱藏

並列摘要

This dissertation studies the data hiding scheme based on the Code Division Multiple Access (CDMA) technique. The row or column taken from the Walsh transformation matrix is used as the spreading code, and the spreading codes are orthogonal to each other, so that the same vector can be embedded overlappingly by multiple secret bits with different spreading codes. Suppose that the secret bits are uniformly distributed, once in the case of multiple embedding with different embedded spreading sequences, most elements of those different spreading sequences are mutually cancelled, so high embedding capacity can be achieved, whereas the image quality will not degrade significantly.
Based on [5], this thesis proposes another embedding sequence and a modified CDMA-based embedding scheme. Next, revises the rhombus-error-prediction method used in [5] to make the resulting prediction error more suitable for the requirement of CDMA-based data embedding, and finally adds the ordering of the prediction errors by the variances of neighboring pixels to continue to enhance the performance of the proposed scheme.
For the lena image, the experimental results show that when compared to the method of [5], the proposed scheme will increase the embedding capacity by 0.0405bpp when the PSNR equals 48.13 dB, and the amount of 0.0075 bpp will be increased when further applying the modification of the diamond-shaped prediction error. Finally, after sorting the diamond-shaped prediction error according to the variances of neighboring pixels, the PSNR can be increased by 0.78 dB under the same embedding capacity of 0.0625 bpp.

並列關鍵字

CDMA ； Reversible data hiding ； Prediction error ； Variance sorting

參考文獻

[1] S. N. Mali, P. M. Patil, R. M. Jalnekar, “Robust and secured image-adaptive data hiding”, Digital Signal Processing, vol. 22, pp. 314-323, 2012.

[2] J. Fridrich, M. Goljan, and Rui Du, “Detecting LSB steganography in color, and gray-scale images,” IEEE Multimedia, vol. 8, no. 4, pp. 22-28, 2001.

[3] Y. Q. Shi, X. Li, X. Zhang, H. T. Wu, and B. Ma, “Reversible data hiding: advances in the past two decades,” IEEE Access, vol. 4, pp. 3210-3237, 2016.

[4] R. C. T. Lee, M. C. Chiu, and J. S. Lin, Communications Engineering. WILEY 2007.

[5] B. Ma and Y. Q. Shi, “A reversible data hiding scheme based on code division multiplexing,” IEEE Trans. On Information Forensics and Security, vol. no. 9, pp. 1914-1927, Sep. 2016.