Logo des Repositoriums

Auflistung nach Autor:in "Drosou, Anastasios"

1 - 2 von 2
  • Konferenzbeitrag
    Activity related biometrics based on motion trajectories
    (BIOSIG 2010: Biometrics and Electronic Signatures. Proceedings of the Special Interest Group on Biometrics and Electronic Signatures, 2010) Drosou, Anastasios; Moustakas, Konstantinos; Ioannidis, Dimos; Tzovaras, Dimitrios
    The current paper contributes to the concept of activity-related biometric authentication in ambient Intelligence environments. The motivation behind the proposed approach derives from activity-related biometrics and is mainly focusing on everyday activities. The activity sequence is captured by a stereoscopic camera and the resulting 2.5D data are processed to extract valuable unobtrusive activity-related features. The novel contribution of the current work lies in the warping of the extracted movements trajectories, so as to compensate for different environmental settings. Au- thentication is performed utilizing both HMM and GMMs. The authentication results performed on a database with 32 subjects show that the current work outperforms existing approaches especially in the case of non-interaction restricting scenarios.
  • Konferenzbeitrag
    Robust 3D face recognition from low resolution images
    (BIOSIG 2013, 2013) Drosou, Anastasios; Moschonas, Panagiotis; Tzovaras, Dimitrios
    This paper proposes a combined approach for robust face recognition from low resolution images captured by a low-budget commercial depth camera. The low resolution of the facial region of interest is compensated via oversampling techniques and efficient trimming algorithms for the generation of an accurate 3D facial model. Two state of the art algorithms for geometric feature extraction are then utilized, i.e. the estimation of the Directional Indices between all the isogeodasic stripes of the same facial surface via the 3D Weighted Walkthroughs (3DW W ) transformation and the estimation of the Spherical Face Representation (SF R). The biometric signature is then enhanced via user-specific cohort biometric templates for each feature, respectively. The experiments have been carried out on the demanding “BIOTAFTOTITA” dataset and the results are very promising even under difficult scenarios (e.g. looking away instances, grimace, etc.). Despite the obvious superiority of the 3DW W transformation over the SF R, it has been noted that the score level fusion of both algorithms improves the authentication performance of the system. On the contrary, only the 3DW W transformation should be preferred in identification scenarios. In- dicatively, the experimental validation on the aforementioned dataset containing 54 subjects illustrates significant succeeds an identification performance of ~ 100% in Rank-1 and Equal Error Rate of 0.25% regarding the authentication performance in the neutral face experiment.