Title:
METHOD AND DEVICE FOR SEPARATING FINGERPRINT FROM OVERLAPPING FINGERPRINT IMAGE BY USING DEEP LEARNING ALGORITHM
Document Type and Number:
WIPO Patent Application WO/2022/131469
Kind Code:
A1
Abstract:
Disclosed are a method and a device for separating a fingerprint from an overlapping fingerprint image by using a deep learning algorithm. A fingerprint separation device: generates a learning data set including a fingerprint overlapping image, in which two or more fingerprints overlap in various backgrounds, and orientation fields and density distributions of multiple artificial fingerprints by using multiple background images and the multiple artificial fingerprints; trains a deep learning network by using the learning data set; and then obtains an orientation field and a density distribution of a fingerprint to be detected, among two or more fingerprints existing in an image to be reconstructed, by using the deep learning network for which the training has been completed, so as to output same to enable fingerprint reconstruction.
More Like This:
Inventors:
LEE BYOUNGHO (KR)
LEE JUHYUN (KR)
CHO JAEBUM (KR)
YOO DONGHEON (KR)
CHAE MINSEOK (KR)
LEE JUHYUN (KR)
CHO JAEBUM (KR)
YOO DONGHEON (KR)
CHAE MINSEOK (KR)
Application Number:
PCT/KR2021/009241
Publication Date:
June 23, 2022
Filing Date:
July 19, 2021
Export Citation:
Assignee:
SEOUL NAT UNIV R&DB FOUNDATION (KR)
International Classes:
G06K9/00; G06N3/08; G06N5/02; G06T3/60; G06T5/00
Foreign References:
| KR20170112302A | 2017-10-12 | |||
| KR20180092197A | 2018-08-17 | |||
| JP2018165911A | 2018-10-25 | |||
| KR20100109752A | 2010-10-11 | |||
| KR20080076170A | 2008-08-20 |
Attorney, Agent or Firm:
Y.P.LEE, MOCK & PARTNERS (KR)
Download PDF: