摘要
针对标准贝叶斯算法在交互建筑设计的应用中还存在效果不好的问题,本文提出了一种基于语义引导贝叶斯算法的交互建筑设计模型,首先针对交互建筑图像某区域从某一角度的理解,采用高斯混合模型实现低层视觉特征到语义的映射,并用概率来表达了对图像对象理解的置信度。然后采用改进贝叶斯算法将中层语义融合,最后用商代法、最小二乘法、微分法等数学方法对运行速度优化。仿真试验结果表明,本文提出的基于语义引导贝叶斯算法的交互建筑设计模型相比较标准贝叶斯,具有更快的运行速度,并且应用效果良好。
There is the problem of ineffective architectural design interactive applications for standardBayesian algorithm, we propose a Bayesian algorithm to guide semantic interaction model based onarchitectural design, the first image of a building for the interaction region from a certain angleunderstanding, using Gaussian mixture model to achieve low-level visual features to semantic mapping,and used to express the confidence probability of image objects to understand. Then the improvedBayesian algorithm middle semantic integration, and finally with Shang method, least squares method,differential method of mathematical methods for optimizing the operating speed. The simulation resultsshow that the proposed algorithm compares the semantic Bayesian interactive guide architectural designstandards-based Bayesian model, with faster speed and with good results.
出处
《科技通报》
北大核心
2016年第5期133-136,共4页
Bulletin of Science and Technology
关键词
贝叶斯算法
交互建筑设计
混合高斯模型
语义提取
运算效率优化
bayesian algorithm
interactive architectural design
gaussian mixture model
semanticextraction
optimizing operational efficiency
作者简介
曲美亭（1982-）,女,汉族,山东,讲师,硕士,艺术设计研究.
