CAJ | 학술논문

针对指纹图像分割的问题，笔者提出了一种基于小波变换突变信号分析原理对指纹边缘进行检测。并结合最小二乘法拟合方法的指纹检测和分割算法。该算法首先将原指纹图像利用平滑函数进行平滑处理，然后使用高斯函数的一阶导数卷积计算水平和垂直方向的突变点，最后使用最小二乘法从不同方向拟合突变点。实验结果表明，该方法能够准确地将指纹图像从背景图像中分割出来。
침대지문도상분할적문제，필자제출료일충기우소파변환돌변신호분석원리대지문변연진행검측。병결합최소이승법의합방법적지문검측화분할산법。해산법수선장원지문도상이용평활함수진행평활처리，연후사용고사함수적일계도수권적계산수평화수직방향적돌변점，최후사용최소이승법종불동방향의합돌변점。실험결과표명，해방법능구준학지장지문도상종배경도상중분할출래。
Aiming at the segmentation of fingerprint image, we proposed a fingerprint detection andsegmentation algorithm which based on the mutations signal analysis of wavelet transformation,combining with theleast square fitting.At first,smoothing is applied to the original fingerprint image by using the smoothing function.Secondly,the mutation points of the horizontal and vertical direction of the fingerprint image is calculated with thefirst derivative convolution of Gaussian function and marked.At last,the least square fitting method is used to fitthe mutation points from different directions.Experiments show that this method can segment the fingerprint imagefrom the background accurately.