基于指纹细节点柱形码的参数自适应选取算法

Parameter adaptive selection method based on fingerprint minutiae cylinder-code

针对现有的细节点柱形编码方法存在柱形编码有效性较差、检索错误率较高的问题,提出了一种改进的细节点柱形编码算法。该算法通过对原有的细节点柱形编码参数进行自适应选取,提高了柱形编码的有效性,有效地降低了指纹检索恢复时的错误率;同时引入局部敏感哈希方法,通过对柱形码进行检索与恢复找回有效的细节点对,提高了匹配的准确率。在FVC2002上的实验结果表明,改进算法具有较好的性能。

In view of the problem of poor cylindrical coding effectiveness and higher retrieval error rate in the existing minutiae cylindrical coding algorithm. This paper circumvented the problem by an improved minutiae cylindrical coding algorithm. It improved the effectiveness of the cylindrical coding and effectively reduced the error rate of the recovery of fingerprint retrieval by adaptive selection of the original cylindrical coding parameters. At the same time, the introduction of a locality sensitive hashing method, during the cylindrical code indexing and recovery, many minutia pairs could be regained and improve the matching accuracy. Experiments performed on FVC2002 database show that the improved algorithm has better performance.