项目名称: 基于冗余字典稀疏表示的置换混叠图像盲分离研究及其应用
项目编号: No.U1204606
项目类型: 联合基金项目
立项/批准年度: 2013
项目学科: 电子学与信息系统
项目作者: 段新涛
作者单位: 河南师范大学
项目金额: 30万元
中文摘要: 盲分离是现代信号处理领域的一个新的研究热点,在诸多领域有着广泛的应用。置换混叠图像盲分离作为一种特殊的单通道盲分离研究的前沿日益受到关注。目前这方面的研究多采用对置换区域的某种操作进行参数估计的方法来检测和分离出置换图像。由于置换混叠图像是由不同来源的图像置换而成,图像类型的多样性和复杂性以及图像在获取处理等环节可能感染噪声等因素,通过估计置换图像所受某种操作的参数往往是不准确的。本项目拟针对不同类型的置换混叠图像,通过设计相应的参数和非参数冗余字典,在稀疏域分解置换混叠图像,检测出置换图像和被置换图像之间在稀疏域的差异,通过差分进化等优化方法分离出置换图像。并应用到插值图像和JPEG图像的篡改图像的盲检测和分离中。
中文关键词: 置换混叠图像;盲分离;稀疏;参数字典;非参数字典
英文摘要: Blind separation is new research direction in modern signal processing field, which has been applied to engineering. There is growing concern on permuted alias image blind separation as frontier of single channel blind separation research. Recent research shows that permuted image can be detected and separated by estimating parameter of permuted image changed with some image processing. However separation results always are inaccurate with these methods above, for those reasons that the permuted alias image is composed of different segments of source images with image diversity and can be noised when thire acquiring and processing phase. This project intends to focus on decomposing permuted alias image in sparse field by devising corresponding parameter and non-parameter redundancy dictionary for various images, detecting the sparse difference between permuting and permuted region, separating the permuted image with optimization methods such as differential evolution. It will be applied to image tamper detection and separation of interpolation and JPEG compressing image.
英文关键词: permuted alias image;blind separation;sparse;parametric dictionary;nonparametric dictionary