Hiding Data In Images Using Least Square Error Mapping

• Main Authors: Ching-Min Hu, 胡敬民
• Other Authors: Ran-Zan Wang
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/47002357048534525864

碩士 === 元智大學 === 資訊工程學系 === 101 === In 2012 Hong and Chen proposed a novel data embedding method using adaptive pixel pair matching (APPM), in which a k-ary(k>=1) number is embedded in each pair of pixels of the cover image. It uses a pixel pair as a reference coordinate in two-dimensional space, and searches for a nearby coordinate in the neighborhood set to replace the pixel pair according to a given message digit. The method not only achieves a high hiding-payload, but also has good resistibility against the attacks of some well-known steganalysis techniques. APPM searches for the best embedding solution in 2D space in which the constraint for the searching space limits the performance of the scheme. In this paper a general form of data embedding using conceptualized coordinate mapping method is proposed. It processes n pixel values of an image block as a reference coordinate in n-dimensional space, and searches for the neighborhood set with least square error to constructs the embedding function. The construction of the proposed scheme with three-pixel block (n=3) under various message bases are illustrated in the paper. Experimental results show that the proposed scheme has higher embedding efficiency than that of the APPM.