现有特征描述方法存在对像素点信息表达不充分的问题.本文提出了一种基于矢量三角形模式的局部特征描述方法,该方法以矢量三角形为基本模式,通过多尺度模式的结合,不局限于提取对称相邻像素点的信息,能更全面地挖掘不同位置像素点之间的信息,并能根据实际应用进行灵活地表达.将这种局部特征描述方法应用于人脸识别中,实验结果表明,基于矢量三角形模式的特征描述及识别方法取得了比LBP等经典算法更精准的效果,证明了该方法的有效性.
The existing pattern description methods often focus on the information between symmetrical and neighboring pixels,thus fail to holistically uncover the implicit nature of images.To alleviate this problem,this paper presents a novel method to describe local features.The method lays its foundation on vector-triangle patterns and mine more comprehensive information of the pixels at different locations with flexible expression according to the requirement of real applications.The extensive experimental results of face recognition have demonstrated that the proposed method achieves better performance than the conventional methods such as LBP.