项目名称: 面向大数据的媒体内容分析与关联语义挖掘研究
项目编号: No.61223003
项目类型: 专项基金项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 瞿裕忠
作者单位: 南京大学
项目金额: 300万元
中文摘要: 在诸如深度Web搜索等应用领域,大数据媒体处理和智能化应用需求日益增多。大数据媒体隐含着很多有价值的信息和语义,且不同来源和类型的媒体数据间隐含着很多深度的信息关联性,因此,分析和挖掘多源异质媒体大数据,将有助于挖掘和发现很多深度信息和事实,因此大数据媒体分析和挖掘技术具有极为重要的研究意义和应用价值。然而大数据媒体分析处理面临很多技术难题。本项目拟研究面向大数据的媒体内容分析与关联语义挖掘技术,重点研究所涉及的复杂的媒体内容和语义分析技术、面向大数据媒体处理的各种机器学习和数据挖掘技术方法、大数据媒体处理并行化计算方法和算法。为了进一步利用多源异质大数据媒体所隐含的丰富的信息关联性完成深度挖掘和事实发现,本项目将在媒体内容和语义分析的基础上,进一步研究多源异质媒体语义的关联融合和深度挖掘技术,为应用行业提供各种媒体挖掘应用和服务。
中文关键词: 语义网;并行计算框架;媒体内容分析;语义融合;机器学习方法
英文摘要: In many business areas such as the deep Web search, more and more demands on big media data processing and intelligent application emerge. The big media data contains a lot of valuable information and semantics and there exists deep information association among different sources and types of media data. Thus analyzing and mining such media data will be very helpful for deep information and fact discovery. Thus big media data analyzing and mining is of great significance for research and application. However, big media data processing comes with great difficulty and challenge. This proposal aims to study on the big media data content analysis and semantic association mining, with focuses on the study of analysis of the media content and semantics, a variety of machine learning methods and techniques towards big media data processing, the big media data parallel computing methods and algorithms. In order to make use of rich information associations hidden in the big media data to perform the deep mining and fact discovery, we will further study the media semantics fusion and deep mining techniques, eventually forming the media mining applications and services for many business areas.
英文关键词: Semantic Web;parallel computing framework;media content analysis;semantic fusion;machine learning method