摘要
马尔可夫随机场模型已成功地应用于计算机低层视觉处理领域,在用MRF建模的过程中引入鲁棒估计算子能增强模型的可靠性,从而提高解的准确性.鲁棒算子在MRF建模中具有双重作用,它不仅能减少观测数据中出格点的影响,而且能应用在信号的平滑约束中,达到根据信号局部特征自适应平滑的效果.最后通过光流场估计的实验验证了作者的结论.
Recently Markov Random Fields (MRF) theory has been applied into many low level vision problems. In this correspondence we argue that if the robust estimators are used in the MRF based modeling, we can improve the reliance of the model and the accuracy of the solution. The robust estimators can reduce the influence of outliers. They also contribute to the discontinuous adaptive signal smoothing techniques. Finally the experiments of optical flow estimation have proved our conclusion.
出处
《东南大学学报(自然科学版)》
EI
CAS
CSCD
1999年第2期100-104,共5页
Journal of Southeast University:Natural Science Edition