Fingerphoto spoofing in mobile devices: A preliminary study
Citations
10 citations
9 citations
Cites background from "Fingerphoto spoofing in mobile devi..."
...However, the attacks at the sensor-level, also known as presentation attacks, can be carried out successfully with utmost ease, as shown for several biometric modalities including face [4], iris [5] and fingerprint [6]....
[...]
7 citations
Cites background from "Fingerphoto spoofing in mobile devi..."
...Similarly, access to ridge-valley details from latent impression of the surface of smartphones [6, 15] or other sources can be used to generate 3D silicone fingerprints [20] (or presentation attack) to spoof the recognition system [21, 26]....
[...]
5 citations
4 citations
Cites background or methods from "Fingerphoto spoofing in mobile devi..."
...Furthermore, in [13], authors have proposed a PAD for fingerphoto recognition using SVM and texture and gradient-based features for print-photo and display attacks where they have achieved an EER of 3....
[...]
...could be used to fool the biometric system [12], [13]....
[...]
...Further, we observe that the proposed scheme shows an improved PAD performance when compared to the state-of-art methods [12], [13], where authors have achieved...
[...]
References
[...]
79,257 citations
"Fingerphoto spoofing in mobile devi..." refers methods in this paper
...Three different matching approaches are adopted: (i) L2 distance based matching, (ii) Neural Network (NN), and (iii) Random Decision Forest (RDF) [6]....
[...]
...We see that the photo attack with iPad-Nokia has the least TAR for ScatNet+NN and ScatNet+RDF matching algorithms, and it is in accordance with Table 3 which provides the highest EER for iPad-Nokia....
[...]
...ScatNet + RDF yields the best results for both spoofed and nonspoofed images; the EERs are in the range of 0.48% to 2.53%....
[...]
40,826 citations
"Fingerphoto spoofing in mobile devi..." refers methods in this paper
...In general, those images that are not correctly matched with the matching algorithm are easily distinguished as spoof images by the SVM classifier, which is in accordance with the basic understanding of the problem....
[...]
...Hence, spoof detection is formulated as a binary classification problem using an SVM [7] to learn these texture patterns from a spoofed image....
[...]
...To study the behavior of high definition display devices such as retina display, gradient based DSIFT [10] features and LUCID descriptor [17] are also independently used to learn an SVM. LUCID descriptors are recently found to provide successful performance in the domain of mobile biometric liveness detection [5]....
[...]
...While using the complete test set, LBP + SVM gives the best spoofed fingerphoto detection performance with 3.71% EER when the complete spoofed dataset is considered....
[...]
...We evaluate the performance of different features such as Local Binary Patterns (LBP), Dense Scale Invariant Feature Transform (DSIFT), and Locally Uniform Comparison Image Descriptor (LUCID) features along with Support Vector Machine (SVM) based fingerphoto spoofing detection algorithm to distinguish between spoofed and non-spoofed images....
[...]
16,989 citations
"Fingerphoto spoofing in mobile devi..." refers methods in this paper
...To study the behavior of high definition display devices such as retina display, gradient based DSIFT [10] features and LUCID descriptor [17] are also independently used to learn an SVM. LUCID descriptors are recently found to provide successful performance in the domain of mobile biometric liveness detection [5]....
[...]
...To study the behavior of high definition display devices such as retina display, gradient based DSIFT [10] features and LUCID descriptor [17] are also independently used to learn an SVM....
[...]
...We evaluate the performance of different features such as Local Binary Patterns (LBP), Dense Scale Invariant Feature Transform (DSIFT), and Locally Uniform Comparison Image Descriptor (LUCID) features along with Support Vector Machine (SVM) based fingerphoto spoofing detection algorithm to distinguish between spoofed and non-spoofed images....
[...]
...The results using both LBP, DSIFT, and LUCID descriptors are presented in Table 5 and Table 6 and the ROC curves are shown in Figure 5....
[...]
...Further, we evaluated different features such as LBP, DSIFT, and LUCID combined with a learning algorithm to classify spoofed and original images....
[...]
5,563 citations
"Fingerphoto spoofing in mobile devi..." refers methods in this paper
...The texture patterns are extracted using LBP features [3, 8, 11]....
[...]
628 citations
"Fingerphoto spoofing in mobile devi..." refers methods in this paper
...The texture patterns are extracted using LBP features [3, 8, 11]....
[...]