基于径向基函数网络的SFS算法研究  

Study of shape from shading algorithm based on the radial basis function network

作　　者：谢明红[1] 

机构地区：[1]华侨大学机电及自动化学院,福建泉州362021

出　　处：《计算机应用》2007年第1期68-70,共3页journal of Computer Applications

摘　　要：分析了现有从明暗恢复形状(SFS)的几种方法普遍存在对恢复的形状的连续性和光滑性的缺点,提出了一种基于径向基函数网络模型进行从明暗恢复形状的新算法。该算法先采用网络构造一个曲面方程,再利用反射函数作为约束条件,通过调整权因子和径向基函数中心和宽度对网络进行自学习,得到一个满意的曲面方程。理论和实验证明,该算法在恢复形状的准确性和曲面的光滑性,连续性上有较大改进。After analyzing the shortcomings of some methods on smoothness and continuity for Shape From Shading （SFS）, a new SFS algorithm based on the radial basis function model was proposed. Firstly it constructed a curved face equation, and then did self-learning by adjusting power gene and center and width of the radial basis function with restriction condition on reflect function, finally gained satisfactory curved face equation. Theoretical analysis and experimental results show that this algorithm improves a lot on veracity of shape recovering and smoothness and continuity of the curved face.

关 键 词：径向基函数网络 从明暗恢复形状 计算机视觉 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]