資訊隱藏（Information Hiding）的技術，使我們能夠在圖像裡嵌入額外的資訊，通常嵌入這些額外的資訊的技術與這些額外的資訊本身的作用有著很大的關連。我們可以根據藏入這些額外資訊的意義，將資訊隱藏的技術分為資料隱藏（Data Hiding）與數位浮水印（Digital Watermarking）兩大領域。在資料隱藏的技術部分，其目標是如何在最不影響圖像品質的條件下，嵌入最多的資訊量，最主要是應用在秘密通訊、資料管理上；而數位浮水印則是考慮到資料的強韌性（robustness），如何在圖像裡嵌入不易讓攻擊者移除的資訊是這個領域的主要課題，數位浮水印的技術通常都是應用在著作權的保護上。
不論資料隱藏或數位浮水印，目前最主要可分成兩個研究領域：一種是將資訊藏在圖像的空間域（Spatial domain）裡，另一種則是將資訊藏在圖像的頻率域（Frequency domain）中。在空間域裡的隱藏方法最常見的方法是將資訊藏在每個pixel的least significant bits(LSB)，而頻率域裡的隱藏方法則是將資訊嵌入在肉眼所不易注意到的頻率帶中。
由於一張圖像同時有空間域與時間域可供資訊隱藏，因此本篇論文提出了一套結合資料隱藏與數位浮水印的方法，在頻率域中利用mean quantization的概念將數位浮水印嵌入到適合的頻帶中，空間域中則以修正後的LSB方法嵌入資料量大的資訊。透過本篇論文提出的機制，可以滿足資料隱藏與數位浮水印的個別需求，進而使圖像兼具著作權保護、秘密通訊與資料管理的應用性。

In order to make one image have an ability to simultaneously cope with copyright protection, secret communication and data management, a novel technique combining data hiding with digital watermarking is presented in this paper.
Firstly, the proposed mechanism embeds the watermark image into the host one on the DCT frequency domain using the concept of mean quantization integrating the human visual system characteristics to achieve more robustness and invisibility. Then it embeds the important information into the covered image on the spatial domain using the modified LSB method to reduce the distortion of the covered image induced by the simple LSB substitution method.
The proposed method has the following advantages : (1) For the digital watermarking, the proposed scheme using the concept of mean quantization improves the robustness of the watermark image but not affects the quality of the host image by means of JND. (2) For the data hiding, the modified LSB substitution method is capable of embedding the same payload the simple LSB method embeds but not inducing the same distortion the simple LSB method induces. (3)Because it does not require the original image in the extraction process, the security can be guaranteed. Furthermore, to enhance robustness, a random permutation of the embedded data is used to defeat the attacks of image processing, such as image cropping.

[1] Frank Y. Shih, Scott Y.T. Wu, "Combinational image watermarking in the spatial and frequency domains", Pattern Recognition 36 (2003) 969-975
[2] Liang-Hua Chen, Jyh-Jiun Lin, “Mean quantization based image watermarking”, Image and Vision Computing 21 (2003) 717-727
[3] Gwo-Jong Yu, Chun-Shien Lu, Hong-Yuan Mark Liao, “Mean-quantization-based fragile watermarking for image authentication”, Optical Engineering, Vol.40 No.7, July 2001 1396-1408
[4] Andrew B. Watson, “DCT quantization matrics visually optimized for individual images”, Human Vision, Visual Processing, and Digital Display IV, SPIE-1913 : 202-216, 1993
[5] Ingemar J. Cox Matthew L. Miller Jeffery A. Bloom; DIGITAL WATERMARKING, p215-p218
[6] Chin-Ching Thien, Ja-chen Lin, “A simple and high-hiding capacity method for hiding digit-by-digit data in images based on modulus function”, Pattern Recognition 36 (2003) 2875-2881
[7] Chin-Shiuh Shieh, Hsiang-Cheh Huang, Feng-Hsing Wang, Jeng-Shyang Pan, “Genetic watermarking based on transform-domain techniques”, Pattern Recognition 37 (2004) 555-565
[8] Chwei-Shyong Tsai, Chin-Chen Chang, Tung-shou Chen, Ming-Huang Chen, “Embedding robust gray-level watermark in an image using DCT transformation”, Distributed multimedia databases: technique & applications, Idea Group Publishing, Hershey, PA, 2002
[9] Chiou-Ting Hsu, Ja-Ling Wu, “Hidden digital watermarks in images”, image processing, IEEE Transactions Vol. 8, No. 1, January 1999