對於壓縮圖像(compressed image)與可逆資訊隱藏(reversible data hiding)的研究而言,在壓縮圖像中隱藏資訊及抽取隱藏資訊後並回覆原壓縮圖像,是其中一個非常重要的目的。Chin-Chen Chang 等人提出了Reversible hiding in DCT-based compressed images的方法,能在壓縮圖像中隱藏資訊,並獲得不錯的實驗結果。不過他們隱藏資訊的方法對圖像品質的影響較為明顯,而且無法提供較高的資訊嵌入量。 在本論文中,我們提出兩種能對壓縮圖像做資訊嵌入、提取與復原的可逆資訊隱藏方法。我們的方法不但計算簡單,程式效率高,而且最後的實驗結果顯示,偽裝圖像(stego-image)擁有良好的視覺品質並具備高負載量(embedding capacity)的能力。
Hiding information in compressed images, extracting the hidden information, and recovering the original compressed images without any distortion are important issues in the study of reversible data hiding and image compression. In previous research, Chin-Chen Chang et al proposed a strategy of reversible data hiding to hide information in DCT-based compressed images. Though the experimental results of their method can provide expected acceptable image quality of compressed stego-images and successfully achieve reversibility, the embedding capacity is not large enough and the image quality can be further improved. In this thesis, we present two strategies of data hiding to hide (extract) information in compressed images. All of the two strategies are reversible. That is, the original compressed images can be recovered without any distortion. Our schemes are easy to implement, computationally efficient. Furthermore, the experimental results show the image quality and the embedding capacity of our schemes are better than Chang’s.