CAJ | 학술논문

为了提高掌纹识别算法识别率，降低特征向量维数，采用基于关键点定位对掌纹进行预处理，利用双向直方图均衡化增强掌纹图像的对比度和纹理细节。使用纹理频谱度量的方法寻找掌纹主线方向，确定Ga-bor滤波器的方向参数，在此基础上，结合多尺度的LBP算子对Gabor变换后的掌纹特征进行提取。最后利用GA优化SVM参数的方法对掌纹图像进行分类。实验结果表明，本文算法能够获得高达97.5%的正确识别率。
위료제고장문식별산법식별솔，강저특정향량유수，채용기우관건점정위대장문진행예처리，이용쌍향직방도균형화증강장문도상적대비도화문리세절。사용문리빈보도량적방법심조장문주선방향，학정Ga-bor려파기적방향삼수，재차기출상，결합다척도적LBP산자대Gabor변환후적장문특정진행제취。최후이용GA우화SVM삼수적방법대장문도상진행분류。실험결과표명，본문산법능구획득고체97.5%적정학식별솔。
In order to improve the recognition algorithms of palmprint recognition rate, reduce the dimension of feature vectors, based on key point location of palmprint preprocessing, using two-dimensional histogram equalization contrast palmprint image and texture detail. Then use the method of texture spectrum measurement for palmprint line direction, determine the direction of the Gabor filter parameters, based on the combination of LBP operator, multi-scale Gabor transform of palmprint features are extracted. Finally, by using the method of GA to optimize the parame-ters of SVM to classify the palmprint image. The experimental results show that, this algorithm can obtain the correct recognition rate up to 97. 5%.