基于连续分布方向图和改进的PoincarIndex的指纹分类-电子学报.PDF

5 Vol. 34 No. 5 2006 5 ACTA ELECTRONICA SIN ICA M ay 2006 PoincarIndex , 2 2 3 2 聂桂军 , 吴 陈, 叶锡君 , 徐荣青 ( . , 225007; 2 , 2 2003; 3. , 2 0095) : (), ; PoincarIndex , Po incarIndex, , ( core delta). , PoincarIn dex 5, ( 4000), , , 9705% , , . : ; ; ; ; Po incarIndex : TP39 : A : 03722 2 ( 2006) FingerprintClassification Based on Both Continuously Distributed Directional Image andModified Version of PoincarIndex , 2 2 3 2 N IE Gu ijun , WU Chen , YE X ijun , XU Rongqing ( . D ep artm ent of C omp uter Sc ience Technology, Yangzh ou Voca tiona l Colleg e of Env irom ent and R e ou rce , Yangzhou, J iang u 225007, Ch ina; 2. The Op en ing L aboratory of Intel ligen t C omp ut ing, Jiang u Un iver ity of Sc ience and T echnology, In titu te of C omp u ting Technology, Chine e A cademy of S cience , Z henj iang, Jiang u 2 2003, Ch ina; 3. S chool of Inf orm a tion S cien ce T echnology, N anj ing A gricu ltura l Un iver ity, N anj ing, Jiang u 2 0095, China) Abstract: A new concept on the continuously distributed directional mi age/field( CDDF) and them ethod to com pute it in the fingerprint mi ages are proposed, which exhibits not only good continuity, w ell gradualness, and excellent ro bustness to the no ises, but very high precision, as well. Then, the classical formula to compute the Po incarIndex and the algorithm for the singularity detection are mi prov