基于Ncut的SIFT特征匹配算法

SIFT Feature Matching Algorithm Based on Ncut

经典尺度不变特征变换(SIFT)特征匹配算法存在实时性差、纹理相似区域易发生误匹配的问题。为此,提出一种基于归一化分割(Ncut)的SIFT特征匹配算法。针对相同背景的运动视频,将归一化分割算法的图论聚类思想融入SIFT特征匹配中,根据运动趋势相似度对特征点进行Ncut运动聚类,再逐类分别匹配,通过缩小各特征点匹配过程中的搜索范围,减少匹配时间及不同特征类之间的误匹配。实验结果表明,该算法能提高匹配效率,对纹理相似区域的误匹配现象有较好的抑制作用,实现了相邻图像帧的特征稳定匹配。

Traditional Scale Invariant Feature Transform（SIFT） feature matching algorithm is based on global search, so bad real-time performance is always shown and mismatches always exist in regions with similar textures. This paper proposes a SIFT feature matching algorithm based on Normalized Cut（Neut）. It puts forward a novel matching method for SIFT feature based on Ncut in the scene of moving objects. It breaks every SIFT feature into its most prominent moving groups and does feature matching respectively between related groups. The proposed algorithm reduces the range of searching and the experimental results show that it provides higher matching efficiency and higher matching accuracy in regions with similar textures. It achieves stable real-time SIFT feature matching between adjacent image frames.