图标形状复杂度的计算度量
发布时间：2018-12-13 08:43
【摘要】：为了解决二维图标形状视觉复杂度的计算度量问题,提出一个基于回归模型的图标形状复杂度计算模型.首先对图标训练数据集进行测试者心理评估;然后对该数据集进行几何特征抽取,并计算得到候选特征变量;最后通过回归分析从候选特征变量中选出4个变量构建回归模型来量化评估图标复杂度.用图标测试数据集对该回归模型进行验证的结果表明,该模型可以解释80%的复杂度人工评估结果;测试数据集的模型量化评估结果和人工评估结果之间斯皮尔曼相关系数达0.922(最大值为1).该模型在图标形状分析、检索、分类等方面具有广泛应用价值.
[Abstract]:In order to solve the problem of calculating the visual complexity of two-dimensional icon shape, a model for calculating the complexity of icon shape based on regression model is proposed. Firstly, the mental evaluation of the icon training data set is carried out, and then the geometric feature extraction is carried out, and the candidate feature variables are calculated. Finally, four variables are selected from the candidate feature variables by regression analysis to construct a regression model to quantitatively evaluate the complexity of icons. The results of verification of the regression model with icon test data set show that the model can explain 80% of the complexity of manual evaluation results. The Spelman correlation coefficient between the quantitative evaluation results of the model and the manual evaluation results of the test data set is 0.922 (the maximum value is 1). The model is widely used in icon shape analysis, retrieval, classification and so on.
【作者单位】： 浙江大学CAD&CG国家重点实验室;浙江大学计算机科学与技术学院;浙江大学教育学院心理系;
【基金】：国家自然科学基金(61772463,61379069) 国家科技支撑计划(2014BAK09B04) 国家社科基金重大项目(12&ZD231)
【分类号】：TP391.41
本文编号：2376262
[Abstract]:In order to solve the problem of calculating the visual complexity of two-dimensional icon shape, a model for calculating the complexity of icon shape based on regression model is proposed. Firstly, the mental evaluation of the icon training data set is carried out, and then the geometric feature extraction is carried out, and the candidate feature variables are calculated. Finally, four variables are selected from the candidate feature variables by regression analysis to construct a regression model to quantitatively evaluate the complexity of icons. The results of verification of the regression model with icon test data set show that the model can explain 80% of the complexity of manual evaluation results. The Spelman correlation coefficient between the quantitative evaluation results of the model and the manual evaluation results of the test data set is 0.922 (the maximum value is 1). The model is widely used in icon shape analysis, retrieval, classification and so on.
【作者单位】： 浙江大学CAD&CG国家重点实验室;浙江大学计算机科学与技术学院;浙江大学教育学院心理系;
【基金】：国家自然科学基金(61772463,61379069) 国家科技支撑计划(2014BAK09B04) 国家社科基金重大项目(12&ZD231)
【分类号】：TP391.41
【相似文献】
相关期刊论文 前4条
1 黄小兰;;比较几种挑选“最优”回归模型的指标[J];中国卫生统计;1988年04期
2 李君;;基于路径的多元标准回归模型及其应用[J];湛江师范学院学报;2006年06期
3 韩立强,姜兴序;机床主轴结构回归模型近似重分析优化设计[J];吉林工学院学报;1990年04期
4 李海涛;;重回归模型在软件开发预估工作量中的应用[J];电脑与电信;2011年03期
相关硕士学位论文 前2条
1 邓健康;基于级联回归模型的人脸配准研究[D];南京信息工程大学;2015年
2 李昊鑫;三维手姿态建模及其交互系统研究[D];电子科技大学;2016年
，本文编号：2376262
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2376262.html
