向量量化編碼(Vector quantization, VQ)概念常被應用到各種不同的領域。邊緣吻合向量量化編碼
(Side-match vector quantization, SMVQ)是一種植基於VQ概念的影像壓方法，它有效的改進了VQ的
壓 效果並且維持其良好的影像視覺品質，在SMVQ的方法其為了要避免預測錯誤所造成的嚴重影像
失真，因此用了標示碼做為辨別區塊是利用VQ 或SMVQ 壓縮的，並確保其能完整無誤的被復原，
而這樣的做法使得SMVQ的壓縮效果打了折扣，為了消除SMVQ所產生的標示碼我們採用了反叢聚
及可還原式資訊隱藏的概念設計新的無標示碼之SMVQ 方法。預期我們所設計出來的方法計算量並
不會比傳統SMVQ高很多。
機密資訊可透過資訊隱藏技術有效地保護並傳送，本研究計畫，我們研究一個利用Sudoku 的新
式可還原原始影像資訊隱藏技術，這個方法將機密資料以9 進制方式表示並將其藏入到負載影像裡，
而其將產生兩張藏有機密的影像，透過對像素對的值進行修改可達到機密資訊的藏入。另外，我們所
設計的方法是具有可還原原始影像的資訊隱藏技術，亦即在機密資訊被提取出來後負載影像也可以被
完整無誤地回復。再者，由於Sudoku 的解非常多，因此應用於本方法中將可提昇傳送機密資訊時的
安全性。
目前已經有很多針對機密資訊傳遞的技術被提出來，在未來的兩年裡，我們將以影像為負載媒體，
分析探討現有的技術，進而提出更高效能的藏入方法，第一年，我們將把資訊隱藏技術應用於提昇邊
緣吻合向量量化編碼的壓縮效果。第二年，我們將利用Sudoku 的特性設計新式的可還原原始影像的
資訊隱藏技術，使機密資訊的傳遞更安全且負載影像在取出機密資訊後可完整地被回復。
The vector quantization (VQ) concept is widely used in many applications. Side-match vector
quantization (SMVQ) is a VQ-based image compression method that offers significantly improved
performance in compression rate while maintaining the image quality of compressed images. To eliminate
distortion propagation, SMVQ requires one extra bit to serve as an indicator identifying which blocks are
encoded by SMVQ or VQ, and to make sure all image blocks can be successfully reconstructed. To eliminate
the indicators generated by SMVQ, we applied declustering and reversible data hiding concept to design our
indicator elimination method. With the indicator elimination method, the computation cost of our method is
not significantly higher than that of SMVQ.
Secret information can be protected by using information hiding techniques. In this project, we
investigate a novel reversible information hiding scheme by using Sudoku. The scheme embeds two secret
digits in the base-9 numeral system into a cover pixel pair by distributing them into two stego pixel pairs at a
time. The scheme achieves higher embedding capacity with acceptable lower visual quality of stego images.
In addition, the scheme is a reversible information hiding scheme in which the original cover image can be
completely recovered after secret data has been extracted. Furthermore, the scheme obtains the security
purpose by sharing secret data into two stego images and using different Sudoku’s solutions for different
secret data transmissions.
Many data hiding techniques have been proposed in the past years. In our project, we will study the
techniques proposed in the recent years and investigating some high performance data hiding schemes in
terms of capacity and visual quality. In first year, we will apply data hiding technique to improve the
performance of SMVQ compression in terms of compression rate. In second year, we will utilize Sudoku to
be a reference table and use it to guild the secret data embedding. The scheme not only achieves secret data
delivery but also can fully reconstruct the cover image after secret data have been extracted.
