基于兴趣点多特征融合的物体识别方法

提出一种基于兴趣点多种特征融合的物体识别方法。利用简化的局部二值模式算子去除Harris冗余角点，提取感兴趣区域的3种特征并加权融合特征，在K最近邻(KNN)方法中引进加权因子计算特征距离函数，得到合适的分类器。实验结果表明，该方法能有效提高物体识别的正确率。

Object Recognition Method Based on Fusing Multi-features of Interest Points

This paper proposes a method for object recognition based on fusing multi-features of interest points. It uses Harris to detect corners, and then uses a simplified Local Binary Patterns(LBP) to wipe out some redundant corners. It gives a weight to every feature according to that it contributes to every class of object. In the K-Nearest Neighbors(KNN), it introduces weight of feature to distance function to achieve a classifier adapted to every object. Experimental results show that this method effectively improves the object recognition accuracy.