多小波融合策略的分段式指纹匹配算法

Segmented fingerprint matching algorithm based on multi-wavelets fusion strategy

针对现有指纹匹配算法准确率低、易受指纹复杂形变影响等缺陷,将多小波理论融合策略与分段式理论相结合,提出多小波融合策略的分段式指纹匹配算法。算法将指纹进行多小波基分解提取特征向量,依据DS证据理论融合得到总特征向量。提出采用归一化局部能量加权互信息匹配度,并以此完成局部初匹配。在此基础上,采用可变大小界限盒法进行全局再匹配。实验结果表明:提出的算法有效提高了算法的准确率,同时也有效避免了指纹复杂形变等因素对指纹匹配算法精度的影响。

Aiming at defects of low accuracy and easy to by affected by complexity of fingerprint deformation of existing fingerprint matching algorithm segmented fingerprint matching algorithm based on multi-wavelets theory fusion strategy is proposed. Combines with multi-wavelets fusion theory and segmented theory. In this method,fingerprint images are decomposed based on multi-wavelets to extract the feature vectors,then the feature vectors is fused into a general feature vector based on Dempster-Shafer( DS) evidential theory. Normalized partial energyweighted-mutual-information matching is introduced as new similarity measures in first-local-matching step. Then changeable sized boundary box method is applied to carry out the second-global-matching step. Experimental results show that the proposed algorithm improves accuracy,while also be slightly influenced by the deformation factor of the fingerprint.