基于MPEG-7标准的人脸结构描述模型的研究

发布时间：2018-07-06 11:15

本文选题：MPEG-7标准 + 人脸结构　；参考：《黑龙江大学》2013年硕士论文

【摘要】：网络多媒体技术已经在我们的生活中占据着主导地位，人们可以从网上获得很多文本和试听信息，它的高速发展，正在逐步改变人们的生产和工作方式。但随之而来的，寻求媒体内容的多样性的需求就显得愈加困难。目前基于文本内容的搜索引擎虽然众多，但这并不满足我们对多媒体信息内容搜索的广泛性、兴趣性和有效性的要求。MPEG-7是对多媒体信息内容描述的重要标准，所以对其的深入发掘研究就具有极大的应用价值。本文主要研究和使用MPEG-7标准构建相关的人脸结构模型，，利用多媒体描述技术研究人脸面部区域的结构和相关内容。在组合人脸图像处理算法的基础上提取五官轮廓和坐标，统一描述人脸图像内容的特征关系，编写相应的索引应用程序，完成图像查询的信息索引。本论文从人脸结构区域的特征关系和图像的标准描述机制入手进行了相关的研究。主要的研究内容可以分为三步：一是大量查阅国内外的相关文献资料，研究MPEG-7的原理和图像处理算法，并利用标准描述方案对图像内容建立完善的描述模型，以实现快速有效的信息检索与定位。二是对人脸数字图像处理的算法进行重组，完成五官分割定位和特征提取，借助图像的颜色直方图分别在频域和空域上对图进行处理，形成丰富的信息互补和信息关联（例如红色的嘴唇、大眼睛）特性，用以满足用户的多条件特征索引需求。三是以人脸结构(主要包括五官和颜色)的整体性为主，结合图像特征与描述定义语言进行全面的MPEG-7人脸描述，通过MPEG-7标准生成多个时间与空间相互关联的文本描述文件，用以提高对图像感兴趣区域内容的快速索引定位。
[Abstract]:The network multimedia technology has already occupied the dominant position in our life, people can obtain a lot of text and audition information from the net, its high speed development, is gradually changing people's production and working way. But with it, the need for diversity in media content becomes increasingly difficult. Although there are many search engines based on text content, this does not meet the requirement of the extensiveness, interest and validity of multimedia information content search. MPEG-7 is an important standard for the description of multimedia information content. Therefore, it has great application value to further explore and study it. This paper mainly studies and uses MPEG-7 standard to construct the related face structure model, and uses multimedia description technology to study the structure and related content of face region. Based on the combined face image processing algorithm, the facial contour and coordinates are extracted, the feature relationship of the face image content is described uniformly, the corresponding index application program is written, and the information index of the image query is completed. In this paper, the feature relation of face structure region and the standard description mechanism of image are studied. The main research content can be divided into three steps: first, a large number of domestic and foreign literature, the principle of MPEG-7 and image processing algorithm, and the use of standard description scheme to establish a complete description model of image content, In order to achieve fast and effective information retrieval and positioning. Secondly, the algorithm of face digital image processing is reorganized to complete facial segmentation and feature extraction, and the color histogram of the image is used to process the image in frequency domain and spatial domain, respectively. Forms rich information complementary and information correlation (such as red lips, big eyes) features to meet the user's multi-conditional feature index requirements. Thirdly, it is based on the integrity of face structure (mainly including facial features and colors), combined with image features and description definition language to carry out a comprehensive MPEG-7 face description, through MPEG-7 standard to generate a number of time and space related text description files. To improve the content of the region of interest to the image fast index location.
【学位授予单位】：黑龙江大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.41

【参考文献】