CAJ | 학술논문

手指指节折痕同指纹和掌纹一样，具有唯一性、稳定性及可区分性的特点，可作为一种用于人体身份识别的生物特征。针对手指指节折痕的分布特征和形状特征，研究一种基于特征融合的指节折痕识别算法。首先利用Radon变换对指节折痕子图像进行变换，形成投影矩阵；其次利用奇异值分解方法对投影矩阵进行奇异值分解，从而形成指节折痕特征矢量；更进一步，考虑指节折痕特征比较简单的特点，重点研究了指节折痕的特征融合策略，用以描述指节折痕特征以达到最大的可区分性；此外定义了一种城区距离来衡量不同指节折痕特征之间的相似度，进行指节折痕特征匹配。最后在自建图像数据库中进行了测试，验证了算法的可行性及有效性。
수지지절절흔동지문화장문일양，구유유일성、은정성급가구분성적특점，가작위일충용우인체신빈식별적생물특정。침대수지지절절흔적분포특정화형상특정，연구일충기우특정융합적지절절흔식별산법。수선이용Radon변환대지절절흔자도상진행변환，형성투영구진；기차이용기이치분해방법대투영구진진행기이치분해，종이형성지절절흔특정시량；경진일보，고필지절절흔특정비교간단적특점，중점연구료지절절흔적특정융합책략，용이묘술지절절흔특정이체도최대적가구분성；차외정의료일충성구거리래형량불동지절절흔특정지간적상사도，진행지절절흔특정필배。최후재자건도상수거고중진행료측시，험증료산법적가행성급유효성。
As well as fingerprints and palmprint,finger crease has characteristics of uniqueness,stability and differentiation,it can be used as a biometric for human identification. According to the distribution and shape feature of finger crease, an algorithm for finger crease recognition based on feature fusion is researched. Firstly, using Radon Transform to transform finger crease sub image, the projection matrix is obtained. Secondly,using the Singular Value Decomposition method to decompose the projection matrix,the feature vector of finger crease is obtained. Furthermore,considering the simplicity of the finger crease,the feature fusion strategy is emphatically researched, which enable combined-feature to have the bigger distinctiveness. In addition, a city distance is defined to measure the similarity between different finger crease feature and the finger crease is matched by the City Block Distance. Finally, the test is performed with the self-built image database and the experimental results demonstrate that the algorithm proposed is feasible and effective.