Application of Canny filter and DWT in fingerprint detection a new approach

Fingerprint detection is one of the most important applications of image processing in a security system. Recent literature deals with fingerprint detection based on pattern recognition, moment based image recognition and discrete wavelet transform (DWT). Among them DWT based analysis requires the least amount of memory space which considers only wavelet coefficient of an image in matching with preserved coefficient. This paper proposes a new approach, which combines image filtering and color inversion in enhancement of contrast along with two dimensional DWT to get the matrix of the image. Finally four statistical parameters: cross correlation coefficient, skewness, kurtosis and convolution of approximate coefficient of one dimensional DWT are compared in detection of fingerprint of a person.