论文部分内容阅读
Recently,fingerprint recognition technique has been used to improve the security of online transaction by smart phone.The size of fingerprint is a key property for a fingerprint recognition system.The larger of image size,the higher of cost and the more accurate of identification.In this paper,we evaluate the relationship between fingerprint image sizes and matching accuracy using a minutia-based fingerprint recognition technique.We computed the equal error rates (EER) on a series of testing databases (TDs) which generate from different sizes of fingerprint images by copying the central regions of larger fingerprint images in database of FVC2004.The experiment results show that if the width and height of a fingerprint image size are greater than some specific values,there won’t be obvious effects on the matching accuracy of the system and when the image width and height are less than some specific values,the matching accuracy decreases abruptly.Finally,we found the suitable image size for fingerprint identification system on smart phones.