摘要
通过对使用V-Ray渲染器的渲染系统进行深入分析,提取出影响渲染时间的13个特征参数,研究了基于粒子群优化支持向量回归机(PSO-SVR)的渲染时间预估方法,采用粒子群算法随机搜索策略优化支持向量回归机的训练参数,获得了较优的支持向量回归机预测模型,实现渲染时间的准确预估.实验结果表明,在渲染时间预估中,PSO-SVR比BP神经网络和逐步回归预测精度高,并且具有较好的泛化能力.
Through in-depth analysis of the rendering system with a V-Ray renderer,we extract the thirteen factors that affect the rendering time,then study time prediction method for rendering based on support vector regression optimized by particle swarm optimization( PSO-SVR). In this method,we adopt random search strategy of particle swarm optimization algorithm to optimize the training parameters of support vector regression,gain the optimized support vector regression forecasting model,and realize the render time accurate forecasts. The experimental results show that in time prediction of rendering,PSO-SVR is much better than BP neural network and stepwise regression,its prediction accuracy is greatly improved,and it has good generalization ability.
出处
《江苏科技大学学报(自然科学版)》
CAS
北大核心
2017年第2期207-213,共7页
Journal of Jiangsu University of Science and Technology:Natural Science Edition
基金
国家自然科学基金资助项目(61572498)
关键词
支持向量回归机
粒子群优化
渲染
渲染时间预估
support vector regression
particle swarm optimization
rendering
render time prediction