基于Contourlet算法的新型图像编码技术研究与验证

发布时间：2018-03-24 06:18

本文选题：Contourlet变换　切入点：图像水印　出处：《吉林建筑大学》2017年硕士论文

【摘要】：随着中国经济高速增长,以物联网和云计算为代表的信息化技术掀起了计算机、通信、信息内容的监测与控制的4C革命,网络功能开始为社会各行业和社会生活提供全面应用。数字电视、遥感图像传输、社交媒体、医用显微图像处理、机器视觉、电子商务等应用领域广泛扩展,使得图像数据越来越庞大,图像的存储和传输成为了日益增长的难题,因此,一种新型的图像编码技术成为了当下最受关注的焦点。本文以Contourlet变换和主成分分析为基础,提出了一种对几何变换和图像鲁棒压缩的一种水印解码器。该方案利用噪声识别功能,自适应调整水印嵌入强度,保持图像的感知质量,在各种可能的失真下可以高精度地检测水印。首先,针对Contourlet变换的原理和性质,本文对Contourlet为基础的图像水印、图像融合、图像压缩领域的应用和发展前景进行了深入探索和介绍,叙述了Contourlet变换的基本原理,为本文研究打下了详实的理论基础。其次,本文利用NIG作为图像的轮廓系数来提出基于NIG分布的最大似然法的水印提取统计方法。使用获得的闭式表达式分别在不存在和存在失真的环境中提取水印位置,以验证所提出解码器的鲁棒性。最后,结合冗余离散小波变换和小波原子变换,对于压缩图像算法提出了一种基于一幅图像中的质地和纹理具有不同的形态特征的方法。通过Contourlet变换和压缩感知的组合重建压缩图像的质地,并通过单层离散小波变换和压缩感知的组合重建压缩图像的纹理。实验结果表明,该方法在保证原始图像主要结构的基础上,确保了低采样率下细节的质量,确保了图像质地和纹理清晰。此外,它还具有较高的压缩率。
[Abstract]:With China's rapid economic growth, the information technology represented by the Internet of things and cloud computing has set off a 4C revolution in the monitoring and control of computers, communications, and information content. The network function began to provide a comprehensive application for various industries and social life. Digital TV, remote sensing image transmission, social media, medical microscopic image processing, machine vision, electronic commerce and other applications have been widely expanded. As the image data becomes larger and larger, the storage and transmission of image becomes a growing problem. Therefore, a new image coding technology has become the focus of attention. This paper is based on Contourlet transform and principal component analysis. In this paper, a watermarking decoder for geometric transformation and image robust compression is proposed, which adaptively adjusts the embedding strength of the watermark and maintains the perceptual quality of the image by using the noise recognition function. The watermark can be detected with high accuracy under various possible distortions. Firstly, according to the principle and properties of Contourlet transform, this paper focuses on the image watermarking based on Contourlet and image fusion. The application and development prospect of image compression are deeply explored and introduced, and the basic principle of Contourlet transform is described, which lays a solid theoretical foundation for the research of this paper. In this paper, NIG is used as the contour coefficient of the image to propose a statistical method of watermark extraction based on the maximum likelihood method of NIG distribution. The obtained closed expression is used to extract the watermark position in the environment where there is no existence or distortion, respectively. Finally, the redundant discrete wavelet transform and wavelet atomic transform are combined to verify the robustness of the proposed decoder. For the compressed image algorithm, a method based on the texture and texture in a single image with different morphological features is proposed. The texture of the compressed image is reconstructed by the combination of Contourlet transform and compression perception. The texture of compressed image is reconstructed by the combination of single-layer discrete wavelet transform and compression perception. The experimental results show that the method ensures the quality of details at low sampling rate on the basis of guaranteeing the main structure of the original image. Ensure the texture and texture of the image clear. In addition, it also has a high compression ratio.
【学位授予单位】：吉林建筑大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【相似文献】