| Summary: | 碩士 === 國立臺北科技大學 === 電腦與通訊研究所 === 96 === Because of the arrival of digital world, media in digital form can be easily altered, duplicated and spread which causes the rights and interests of original author and copyright owner are compromised or destroyed. Therefore, the protection of the intellectual property rights of digital information has come to attention. Digital watermark provides a simple and effective way of protection for property of the kind. Nowadays, more and more studies have been thrown to the field of digital watermark.
In this thesis, a method of robustness and blind extraction watermark has been put forward. The main purpose of this method is to protect the intellectual property right of static images. The basic concept is to utilize wavelet transformation where, according to different characteristic coefficient on different band, three methods will be applied respectively as following.
1. Lattice code based on communication principle
2. Insignificant coefficient based on JND of human visual model
3. Quantization Index Modulation based on SVD
The three methods above will achieve the goal of embedding watermark on the premise of keeping image fidelity.
The experimental results prove that the proposed method presented in this thesis will have immunity to the attacks of various image compression, commonly used image processing methods and geometric methods. Additionally, because this approach is blind watermark, which means neither original image and nor its related information is needed, it is a very convenient and practical method of watermark.
|
|---|
