中文摘要：

由于指纹的残缺区域缺失有效信息，使得通过细节点和脊线信息来进行指纹匹配的工作难以进行．本文针对指纹残缺区域中脊线断裂和细节点缺失的现象，提出了一种重建算法．融合邻域中的细节点和方向场信息对残缺区域的方向场进行估计；然后基于估计出的方向场并结合先验知识来分析残缺部分的脊线与细节点分布情况，对其进行修复；最后，引入信息熵对多种重建方案进行评估，从而确定残缺部分的最佳重建结果．实验结果表明，该算法能够对面积较大并可能包含细节点的残缺区域进行较好的重建，进而提高后续匹配过程的准确率．

英文摘要：

An incomplete fingerprint is one which contains noise and cannot be recognized in some areas.In some cases,it may not provide sufficient minutiae or ridge details for undertaking a normal matching process. This paper proposes an algorithm to reconstruct incomplete regions. The nearest minutiae and orientation field of neighborhood are combined to estimate the unknown orientation field. Previous outcome is used to connect disrupted ridges and recover missed minutiae based on ridge topology features, and arrive several rational schemes. They are measured by using entropy and the best is chosen. The experimental results demon strate that proposed algorithm is effective in recomlrucling incomplete fingerprint, and hence improve the performance of matching.