在資訊隱藏的研究領域中，提高藏密量亦是一項重要的研究議題．有些學者利用一種稱為向量量化的壓縮方法，將機密影像進行壓縮後，將所產生的索引表藏入向量量化的編碼簿中，再將藏有機密影像的編碼簿傳送到接收端．接收端將索引表取出後，再利用此編碼簿還原機密影像，然而，此時所使用的編碼簿由於已藏有索引表，因此與原先用來壓縮機密影像的編碼簿有所出入，使得所還原的機密影像品質並不是太好．提高傳送端所使用的編碼簿品質，可以讓還原的機密影像品質有少許的提昇，然而效果有限．在本研究計畫中，我們將提出一套改進的方法，能夠架構在高品質的編碼簿上，再更提昇還原的機密影像品質．我們的方法主要的概念是，將編碼簿的編碼字進行重排，使得機密影像壓縮後所產生的索引表，其位元串能與編碼簿的最低位元平面更接近，因此索引表藏入編碼簿後，對編碼簿造成的破壞更小，進而可提高機密影像的還原品質．由於一本編碼簿所包含的編碼字非常多，因此可能的重排方式非常多種，本研究計畫將應用基因演算法來搜尋最佳的重排方式．
One of important criteria of an image hiding scheme is the hiding capacity. To increase the hiding capacity, some researchers try to employ vector quantization (VQ) to compress the secret image and then embed the index table, i.e. the compressed result, into the codebook. Obviously, the codebook is modified after the secret is embedded. In the receiver site, the secret is recovered with the modified codebook; hence the quality of the recovered secret image may not be as good as the original one. Using a good quality codebook is a possible solution to improve the quality of the recovered secret image, but the improvement is limited. The aim of this research is to propose a modified VQ-based image hiding scheme, which can improve the quality of the recovered secret image more. The main idea is to permute the order of codewords of the codebook, so that the bitsream of the index table can match the least significant bits of the codebook much more. Since the possible permutations are very large, we adopt genetic algorithm to find an optimal permutation in the proposed scheme.
