http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 주제어 : fingerprint liveness detection (검색결과 4 건)
- 무료
- 기관 내 무료
- 유료
-
Convolutional Neural Networks 특징을 이용한 지문 이미지의 위조여부 판별 및 시각화
김원진(Weon-jin Kim),이경수(Qiong-xiu Li),박은수(Eun-soo Park),김정민(Jung-min Kim),김학일(Hak-il Kim) 한국정보보호학회 2016 정보보호학회논문지 Vol.26 No.5
최근 지문 인식을 통한 사용자 인증 기술이 상용화 되면서 위조 지문 이미지 판별이 더욱 중요해졌다. 본 논문에서는 CNN 특징을 이용한 위조 지문 이미지 판별 방법을 제안하였으며, CNN 모델이 실제 지문의 어느 부분에 반응하여 위조지문을 분류하는지 시각화 방법을 통해 분석하였다. 제안하는 방법은 지문영역과 배경영역을 분리하는 전처리 작업후 CNN 모델을 이용하여 지문의 위조여부를 분류한다. 지문을 단순히 생체지문과 위조지문으로 분류하는 것이 아니라 위조지문을 구성하는 물질별로 분류하여 생체지문과 위조지문들에 대한 특징분석을 제공한다. 실험에 사용한 데이터베이스로는 생체 지문 이미지 6500여 장과 위조 지문 이미지 6000여 장으로 구성되어 있는 LivDet2013을 사용하였으며 위조여부에 대한 ACE 값으로 3.1%, 구성 물질 분류 정확도는 평균 79.58%를 보여 높은 수준의 분류성능을 갖고 있음을 확인하였다. With the growing use of fingerprint authentication systems in recent years, the fake fingerprint detection is becoming more and more important. This paper mainly proposes a method for fake fingerprint detection based on CNN, it will visualize the distinctive part of detected fingerprint which provides a deeper insight in CNN model. After the preprocessing part using fingerprint segmentation, the pretrained CNN model is used for detecting the liveness detection. Not only a liveness detection but also feature analysis about the live fingerprint and fake fingerprint are provided after classifying which materials are used for making the fake fingerprint. Our system is evaluated on three databases in LivDet2013, which compromise almost 6500 live fingerprint images and 6000 fake fingerprint images in total. The proposed method achieves 3.1% ACE value about the liveness detection and achieves 79.58% accuracy on LiveDet2013.
-
Fingerprint Liveness Detection Using Gray Level Co-Occurrence Matrix Based Texture Feature
Chengsheng Yuan,Zhihua Xia,Xingming Sun,Decai Sun,Rui Lv 보안공학연구지원센터 2016 International Journal of Grid and Distributed Comp Vol.9 No.10
Fingerprint-based recognition systems have been widely deployed in numerous civilian and government applications. However, the fingerprint recognition systems can be deceived by using an accurate imitation of a real fingerprint such as an artificially made fingerprint. In this paper, we propose a novel software-based fingerprint liveness detection algorithm based on gray level co-occurrence matrix (GLCM), from which we can calculate the texture features of fingerprint images and obtain satisfactory results. For the first time, we extract texture features by constructing four-direction GLCMs in an image, and then quantization operation and normalization operation are adopted. After these, we detected whether a fingerprint image belongs to a real fingerprint or an artificial replica of it. A trained RBF SVM (support vector machine) classifiers scheme is used to make the final live/spoof decision via training and testing feature vectors. The experimental results reveal that our proposed method can discriminate between live fingerprints and fake ones with high classification accuracy.
-
Fingerprint Liveness Detection Using Difference Co-occurrence Matrix Based Texture Features
Zhihua Xia,Chengsheng Yuan,Xingming Sun,Rui Lv,Decai Sun,Guangyong Gao 보안공학연구지원센터 2016 International Journal of Multimedia and Ubiquitous Vol.11 No.11
Fingerprint authentication systems have been widely deployed in both civilian and government applications, however, whether fingerprint authentication systems is security or not has been an important issue under fraudulent attempts through artificial spoof fingerprints. In this paper, inspired by popular feature descriptors such as gray level co-occurrence matrix (GLCM) and Gradient (difference matrix (DM)), we propose a novel software-based fingerprint liveness detection algorithm called difference co-occurrence matrix (DCM). In doing so, quantization operation is firstly conducted on the images. DMs are constructed by calculating difference matrices of horizontal and vertical pixel values of images; difference co-occurrence arrays are constructed from the difference matrices between adjacent pixels. To reduce the influence of abnormal pixel values, truncation is used for DMs. Then, we compute four parameters (Angular Second Moment, Entropy, Inverse Differential Moment and Correlation) used as feature vectors of fingerprint images. For the first time in the fingerprint liveness detection, we construct eight difference co-occurrence matrices and extract texture features from processed DCMs. Finally, SVM classifier is used to predict classification accuracy. The experimental results reveal that our proposed method can achieve more accurate classification compared with the best algorithms of 2013 Fingerprint Liveness Detection Competition, while being able to recognize spoofed fingerprints with a better degree of accuracy.
-
패치기반 컨볼루션 뉴럴 네트워크 특징을 이용한 위조지문 검출
박은수(Eunsoo Park),김원진(Weonjin Kim),이경수(Qiongxiu Li),김정민(Jungmin Kim),김학일(Hakil Kim) 한국정보보호학회 2017 정보보호학회논문지 Vol.27 No.1
최근 모바일 기기에서의 생체인증 시스템의 증가와 출입관리 시스템에서의 위조지문을 이용한 출입 기록 조작으로 인해 위조 지문 검출에 대한 논의가 다시 활발해지고 있다. 본 논문에서는 입력 지문영상을 패치들로 나누고, 각 패치들에 CNN을 적용하여 위조, 생체, 배경의 세 가지로 분류한다. 이 중 배경으로 분류된 패치들을 제외하고 위조와 생체로 분류된 패치들의 수를 세어서 더 많은 패치가 인식된 쪽으로 위조여부를 판단하게 된다. CNN에 배경 클래스를 추가하여 분류하기 때문에, 제안하는 방법은 영상분할과 같은 추가적인 전처리 과정이 필요하지 않다. 제안하는 방법은 LivDet2011, LivDet2013, LivDet2015에 대하여 실험을 진행하였으며 분류결과 3.06%의 평균 오검출을 보여 매우 우수한 성능을 나타냄을 확인하였다. Nowadays, there have been an increasing number of illegal use cases where people try to fabricate the working hours by using fake fingerprints. So, the fingerprint liveness detection techniques have been actively studied and widely demanded in various applications. This paper proposes a new method to detect fake fingerprints using CNN (Convolutional Neural Ntworks) based on the patches of fingerprint images. Fingerprint image is divided into small square sized patches and each patch is classified as live, fake, or background by the CNN. Finally, the fingerprint image is classified into either live or fake based on the voting result between the numbers of fake and live patches. The proposed method does not need preprocessing steps such as segmentation because it includes the background class in the patch classification. This method shows promising results of 3.06% average classification errors on LivDet2011, LivDet2013 and LivDet2015 dataset.
ScienceON
DBpia연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
