项目名称: 绘画鉴藏中的交互式演化算法图像分析研究
项目编号: No.61300127
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 刘罡
作者单位: 湖北工业大学
项目金额: 23万元
中文摘要: 绘画鉴藏通过对作品的线条、色彩、构图等形式因素进行感受分析,对作品的艺术价值进行评判。绘画鉴藏一般依赖人的主观评判,因此会造成结论的不可靠导致经济上的损失。交互式演化算法应用于绘画鉴藏,将在很大程度上克服这种主观性,为绘画鉴藏提供科学依据。本项目研究基于交互式演化算法的图像聚类和特征提取方法。(1)研究基于交互式演化算法的图像聚类方法,对单幅绘画作品进行分割,提取绘画形式因素用于分析,对多幅绘画作品进行聚类分析,用于绘画作品比较;(2)研究基于交互式演化算法的图像特征提取方法,为多幅绘画作品聚类提供基础并建立特征库;(3)针对本项目中的应用,研究交互式演化算法的去噪方法,混合优化问题处理及多人在线评价方法,设计符合实用的交互式演化算法;(4)建立基于交互式演化算法的在线绘画作品鉴藏平台,并做应用推广。在绘画鉴藏中采用图像分析方法是计算机科学与人文学科结合的新领域,有重要理论意义和应用价值。
中文关键词: 交互式演化算法;图像聚类;图像分类;混合优化;深度学习
英文摘要: Painting collection is to make the judgment of artistic value through feeling and analysing the lines, color, composotion and other form factors of the work. Painting collection generally relies on the subjective judgment of the people, and therefore it can create unreliable conclusions to cause economic losses. Interactive evolutionary computation applied to painting collection will largely overcome this subjectivity,and will provide the scientific basis for painting collection. The project to research interactive evolutionary computation-based image clustering and feature extraction methods.(1)This project researches interactive evolutionary computation-based image clustering. It is used to segment a signal work to extract the form factors of the painting for analysis,and to cluster multiple paintings for comparison of paintings;(2)This project researches interactive evolutionary computation-based image feature extraction method.It provides the basis of multiple paintings clustering and establishes a feature library;(3)For the application of the project, it researches the denoising methods, hybrid optimization problems processing and the multiplayer online evaluation methods of interactive evolutionary computation, and then designes to meet the practical interactive evolutionary computation;(4)This project e
英文关键词: Interactive evolutionary computation;Image clustering;Image classification;Hybrid optimization;Deep learning