节点文献
基于BP改进算法的自由曲面重建
Freeform Surface Reconstruction Based on Adaptive BP Algorithm
【摘要】 针对BP基本算法应用于曲面重建的缺点,提出自调整S型函数方法,并利用共轭梯度法同时调整权值系数和S型函数的形状参数,使得收敛曲线能够迅速摆脱平缓的区域,加速了收敛进程。将这一改进的算法应用于BP网重建自由曲面中,仿真结果表明,同BP基本算法相比,在收敛速度和重建精度上均有较大程度的提高。
【Abstract】 This paper proposed a novel algorithm for freeform surface reconstruction using self-adjust Sigmoid function based on the analysis of convergence of the conventional BP algorithm. It uses conjugate gradient algorithm to adjust the weights and coefficients of the shape of Sigmoid function. It makes the convergence curve break away from flat area rapidly, and accelerate the convergence. The advanced arithmetic is used in freeform reconstruction and the result of emulation shows that the improved arithmetic makes considerable progress both in convergence speed and the reconstruction accuracy in contrast to conventional BP algorithm.
【Key words】 Freeform Surfaces; BP Neural Networks; Conjugate Gradient; S Function;
- 【文献出处】 电脑知识与技术(学术交流) ,Computer Knowledge and Technology(Academic Exchange) , 编辑部邮箱 ,2007年06期
- 【分类号】TP18;TP391.7
- 【被引频次】4
- 【下载频次】103