Logo des Repositoriums

Auflistung nach Autor:in "Yamada, Shigefumi"

1 - 6 von 6
  • Konferenzbeitrag
    A biometric key-binding scheme using lattice masking
    (BIOSIG 2014, 2014) Sugimura, Yuka; Yasuda, Masaya; Yamada, Shigefumi; Abe, Narishige; Shinzaki, Takashi
    Template protection technology can protect the confidentiality of a biometric template by certain conversion. We focus on the key-binding approach for template protection. This approach generates a secure template (or a conversion template) from joint data of a user's specific key with a user's template, and the key can be correctly extracted from the secure template only when a queried biometric feature is sufficiently close to the original template. While almost all conventional schemes use the error correcting code (ECC) technique, we present a new technique based on lattices to give a new key-binding scheme. Our proposed scheme can provide several requirements (e.g., diversity and revocability) for template protection, which cannot be provided by ECC-based schemes such as the fuzzy commitment and the fuzzy vault.
  • Konferenzbeitrag
    Evaluation of independence between multiple fingerprints for multibiometrics
    (BIOSIG 2014, 2014) Yamada, Shigefumi; Shinzaki, Takashi
    Multibiometrics provides high recognition accuracy and population coverage by combining different biometric sources. However, some multibiometrics may obtain smaller-than-expected improvement of recognition accuracy if the combined biometric sources are dependent in terms of a false acceptance by mistakenly perceiving biometric features from two different persons as being from the same person. In this paper, we evaluated whether or not features of multiple fingerprints from a same person are statistically independent. By evaluating false acceptance error using matching scores obtained by Verifinger SDK, we confirmed that these features were dependent and the FAR obtained by a fusion of the multiple fingerprints could be affected by the dependence.
  • Konferenzbeitrag
    Evaluation of independence between palm vein and fingerprint for multimodal biometrics
    (BIOSIG 2012, 2012) Yamada, Shigefumi; Endoh, Toshio; Shinzaki, Takashi
    Multimodal biometrics provides high recognition accuracy and population coverage by combining different biometric sources. However, some multimodal biometrics may obtain smaller-than-expected improvement of recognition accuracy if the combined biometric sources are dependent in terms of a false acceptance by mistakenly perceiving biometric features from two different persons as being from the same person. In this paper, we propose our multimodal biometric prototype that captures a palm vein and three fingerprints simultaneously and we evaluate whether or not their combination is statistically independent. By evaluating false acceptance using the palm vein images and the fingerprint images collected with our prototype, we confirmed that the combination of the palm vein and the fingerprints is almost independent.
  • Konferenzbeitrag
    Evaluation on Biometric Accuracy Estimation Using Generalized Pareto (GP) Distribution
    (BIOSIG 2021 - Proceedings of the 20th International Conference of the Biometrics Special Interest Group, 2021) Yamada, Shigefumi; Matsunami, Tomoaki
    The accuracy of biometric authentication technology is becoming more sophisticated with its progress. For this reason, a huge number of biometric samples are required for accuracy evaluation, and the increased collection cost is an issue for biometric vendors. This work establishes a biometric accuracy estimation method using an extreme value theory to reduce the collection cost. It also explains the estimation procedure of false match rate using the generalized Pareto distribution and shows results applied to the face, gait, and voice comparison score data with an estimation effect of about 5–10 times. We investigate the criteria for the applicability of extremum statistics through application cases.
  • Konferenzbeitrag
    Learning by Environment Cluster s for Face Presentation Attack Detection
    (BIOSIG 2021 - Proceedings of the 20th International Conference of the Biometrics Special Interest Group, 2021) Matsunami, Tomoaki; Uchida, Hidetsugu; Abe, Narishige; Yamada, Shigefumi
    Face recognition has been used widely for personal authentication. However, there is a problem that it is vulnerable to a presentation attack in which a counterfeit such as a photo is presented to a camera to impersonate another person. Although various presentation attack detection methods have been proposed, these methods have not been able to sufficiently cope with the diversity of the heterogeneous environments including presentation attack instruments (PAIs) and lighting conditions. In this paper, we propose Learning by Environment Clusters (LEC) which divides training data into some clusters of similar photographic environments and trains bona-fide and attack classification models for each cluster. Experimental results using Replay-Attack, OULU-NPU, and CelebA-Spoof show the EER of the conventional method which trains one classification model from all data was 20.0%, but LEC can achieve 13.8% EER when using binarized statistical image features (BSIFs) and support vector machine used as the classification method
  • Konferenzbeitrag
    A novel local feature for eye movement authentication
    (Biosig 2016, 2016) Abe, Narishige; Yamada, Shigefumi; Shinzaki, Takashi