基于显著图的SIFT特征检测与匹配

基于尺度不变特征变换(SIFT)特征的图像匹配存在特征点数量大、运算时间长等问题。为此，引入视觉注意机制，提出一种基于显著图的SIFT特征检测与匹配方法。比较常用的显著图计算模型，选择谱残差方法提取图片的显著图。对显著图进行二值化和形态学等处理，得到规则合理的显著区域。在显著区域内提取SIFT特征，生成特征向量，进行图像匹配。实验结果表明，该方法能提高运算效率，并且得到的SIFT特征更加稳定。

SIFT Feature Detection and Matching Based on Salient Map

Image matching based on Scale Invariant Feature Transform(SIFT) feature is time-consuming,and there are always a large number of feature points.By introducing salient map into feature extraction and image matching,a new SIFT feature detection and matching method is put forward.Salient computing models are compared,and an efficient spectral residual method is selected to compute the salient map.Then the salient map is processed with binarization and morphology to get regular salient area.SIFT features are detected and matched just in the salient area instead of in the whole image.Experimental results show that the proposed method is much faster and the features in salient area are more stable.