The Watermarking Algorithm Of Image Being Embedded Into Audio Based On Digital Halftone Technology

Audio digital watermarking technology is based on audio as the carrier. The traditional embedded audio digital watermarking is based on binary images primarily, low capacity and less information. In this thesis, we firstly proposed a watermark algorithm embedding gray-scale image into audio based on the image digital halftoning and inverse halftoning technology. Then we further proposed a watermark technology imbedding color image into audio based on the color image’s color space and digital halftoning technology. Color image being watermark greatly improves the watermark embedding capacity. In order to enhance the security of audio watermark, image scrambling technology is adopted. The prime work in this thesis is as follows:1. Introduce the digital watermark technology basic concept and the basic characteristic, as well as digital watermark classification. Analyze the Digital watermark technology research purposes and significance.2. Introduce the basic model and algorithm of audio watermarking technology in detail. The basic algorithm consists mainly of time-domain algorithm and frequency domain algorithms. Describe the evaluation criteria of audio digital watermarking and attack methods.3. Describe the basic knowledge of digital images, analysis of the digital halftone technique and inverse halftoning technology. Implement the digital halftoning and inverse halftoning algorithms.4. Propose gray-scale image audio watermark algorithms based on halftoning technology. In the algorithm, a gray image is binaried by image halftoning. Comparing with the traditional gray-scale image8bits embedding directly, it greatly reduces the capacity of embedded data.5. Propose color image audio watermark algorithm based on the color image’s color space and digital halftoning technology in this thesis. Comparing with grayscale images, color images have more information and better sensory quality. Experimental results show that color images watermark algorithms have high embedding capacity, robustness and high correct rate of extraction.