基于小波矩不变量的模式识别方法

PATTERN RECOGNITION BASED ON WAVELET MOMENT INVARIANTS

为了得到一组局部最优的特征组 ,提出一种结合分散度和顺序前进法 (SFS)的特征选择方法 .实验结果表明小波矩不变量比 Zernike矩不变量有更好的识别效果 ,尤其对于形体相似的物体 。

To get a locally optimum feature set, a feature selection method was presented, which combines divergence with Sequential Forward Selection (SFS). Experimental results prove that wavelet moment invariants are superior to Zernike’s moment invarinats for pattern recognition, especially for classifying seemingly similar objects with subtle difference.