多视点识别三维运动目标新方法

THE NEW METHOD OF RECOGNIZING THREEDIMENSIONAL\=MOVING OBJECT FROM MULTIPLE VIEW

本文讨论了从多视点识别三维运动目标问题．该方案基于对连续输入的二维图像聚类，构成三维目标的特征视图和转移矩阵，利用极－对数坐标变换（ＬＰＭ）和离散付立叶变换（ＤＦＴ）提取出与目标二维特征视图的位置、比例和旋转无关的特征向量．ＡＲＴ－２模型作为目标特征信息的存储器和分类器．实验中对ＡＲＴ－２神经网络进行了改进，取得了满意的结果．

This paper addresses the problem of recognizing threedimensional(3D) moving object from multiple views. It is based on the 2D processed frames of a video sequence which are clustered into view categories called feature aspects of the object and their transitions. Logpolar mapping(LMP) and Discrete Fourier Transformation(DFT) are used for getting the position, scale and rotation invariant feature vectors of 2D characteristic views. ART2 model is used as memory and classifier of the feature information of the object. Improved ART2 neural network is used in experiment, and the results are satisfactory.