A Data Hiding Technique Based on Statistical Analysis

• Main Authors: Chen Hsin Li, 李振祥
• Other Authors: 洪國寶
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/01060670615951092569

碩士 === 國立中興大學 === 資訊科學與工程學系所 === 98 === This thesis presents a data hiding technique modified from the Sudoku-based scheme presented by Chang et al. The proposed method introduces a new method for adapting the embedding positions in the corresponding matrix to improve the visual quality of the stego image. We first count up the frequencies of the secret message preparing to be embedded in the corresponding block. Next, the embedding positions can be adapted according to the sorted order of the frequencies. That is, the secret number with the largest appearance frequency is embedded in the position where the difference between the pixel pair is minimal. Experimental results show that the proposed method is efficient and, in average, the visual quality of the stego image is 1.83dB higher than the previous work. Moreover, we combine the proposed scheme with the Minimal Euclidean Distance searching method presented by Hong et al. The experiments show that the visual quality of the stego image is improved about 0.87dB in average. This result demonstrates that the proposed scheme combined with the modified searching method in this work is efficient and stable. Finally, we modify the reference table presented by Chang by changing it into a 4x4 table. The data size in this table is so smaller than the original reference table that it is more favorable to exchange between the transmitter and the receiver. By experiments, the visual quality of the stego image is significantly improved, if the embedded data with high repetitiveness are used.