基于卷积稀疏表示的内外部指纹融合方法研究

Research on internal and external fingerprint fusion method based on convolutional sparse representation

针对多传感器采集的内外指纹质量各有优缺点的问题,提出结合方向确定度(OCL)质量评估和基于卷积稀疏表示的形态成分分析模型(CSMCA)的内外指纹融合方法OCL-CSMCA。通过将内部与外部指纹图像进行分解、添加像素级方向确定度(OCL)约束、融合与重建,实现内外指纹图像融合。对比实验表明,本文融合指纹图像在改善视觉效果、提高指纹质量、增大有效面积、指纹匹配能力以及细节点正确提取等方面表现良好。

Internal and external fingerprints collected by multiple sensors have their respective advantages and disadvantages.A new internal and external fingerprint fusion method called OCL-CSMCA is proposed,which combines the orientation certainty level(OCL)evaluator and the convolutional sparse representation based morphological component analysis model(CSMCA).By internal and external fingerprint images decomposition,pixel-level OCL constraints integration,fusion and reconstruction,the internal and external fingerprint fusion is realized.Comparative experimental results show that the proposed method can better improve the visual effect and quality of the fused fingerprints,while increasing the effective area.In addition,the fused fingerprints obtained by the proposed method perform well in terms of minutiae extraction and matching.