Logo des Repositoriums

Auflistung nach Schlagwort "Finger Vein Recognition"

1 - 4 von 4
  • Konferenzbeitrag
    Cross-Sensor Finger Vein Recogition
    (BIOSIG 2022, 2022) Bernhard Prommegger, Georg Wimmer and Andreas Uhl
    If biometric systems are rolled out on a large scale, it will not always be guaranteed that all capturing devices are of exactly the same type. It is therefore important to ensure that the biometric system also works across multiple capturing devices. In finger vein biometry, there is almost no published work in this regard. The biggest problem here is certainly that there are only very few datasets (recorded by different institutions) that usually have no overlap at all with the test persons contained. In a first approach, this article tries to examine how well cross-sensor finger vein recognition works. For the investigation, four publicly available datasets, which were acquired with four different devices in three different scenarios, were evaluated. Using three different finger vein recognition approaches, we will show that the results distinctly deteriorate in cross-sensor recognition scenarios compared to the recognition results using only images from the same device, even more so for image data from contact-less and contact-based capturing devices
  • Konferenzbeitrag
    Perspective Multiplication for Multi-Perspective Enrolment in Finger Vein Recognition
    (BIOSIG 2019 - Proceedings of the 18th International Conference of the Biometrics Special Interest Group, 2019) Prommegger, Bernhard; Uhl, Andreas
    Finger vein recognition deals with the identification of subjects based on their venous pattern within the fingers. It has been shown that its recognition accuracy heavily depends on a good alignment of the acquired samples. There are several approaches that try to reduce the impact of finger misplacement. However, none of this approaches is able to prevent all possible types of finger misplacements. As finger vein scanners are evolving towards contact-less acquisition, alignment problems, especially due to longitudinal finger rotation, are becoming even more important. One way to tackle this problem is capturing the vein structure from different perspectives during enrolment, but cost and complexity of capturing devices increases with the number of involved cameras. In this article, a new method to reduce the number of cameras needed for multi-perspective enrolment is presented. The reduction is achieved by introducing additional pseudo perspectives in-between two adjacent cameras. The obtained perspectives are used for additional comparisons during authentication. This way, the complexity of the enrolment devices can be reduced while keeping the recognition performance at a high level.
  • Konferenzbeitrag
    The Two Sides of the Finger - An Evaluation on the Recognition Performance of Dorsal vs. Palmar Finger-Veins
    (BIOSIG 2018 - Proceedings of the 17th International Conference of the Biometrics Special Interest Group, 2018) Kauba, Christof; Prommegger, Bernhard; Uhl, Andreas
    Vascular pattern (vein) based biometrics, especially finger- and hand-vein recognition gain more and more attention. In finger-vein recognition, the images are usually captured from the palmar (bottom) side of the finger. Dorsal (top) side finger vein recognition has not got much attention so far. In this paper we establish a new, publicly available, two-sided (dorsal and palmar) finger-vein data set. The data set is captured using two custom designed finger vein scanners, one based on near-infrared LED illumination, the other one on near-infrared laser modules. A recognition performance comparison between the single subsets (palmar and dorsal) as well as cross-subset (palmar vs. dorsal) comparison is conducted using several well-established finger-vein recognition schemes. The experimental results confirm that the palmar side achieves the overall best recognition performance but in general the dorsal side works better due to inherent finger texture information.
  • Konferenzbeitrag
    Vein Enhancement with Deep Auto-Encoders to improve Finger Vein Recognition
    (BIOSIG 2021 - Proceedings of the 20th International Conference of the Biometrics Special Interest Group, 2021) Bros, Victor; Kotwal, Ketan; Marcel, Sébastien
    The field of Vascular Biometric Recognition has drawn a lot of attention recently with the emergence of new computer vision techniques. The different methods using Deep Learning involve a new understanding of deeper features from the vascular network. The specific architecture of the veins needs complex model capable of comprehending the vascular pattern. In this paper, we present an image enhancement method using Deep Convolutional Neural Network. For this task, a residual convolutional auto-encoder architecture has been trained in a supervised way to enhance the vein patterns in near-infrared images. The method has been evaluated on several databases with promising results on the UTFVP database as a main result. In including the model as a preprocessing in the biometric pipelines of recognition for finger vein patterns, the error rate has been reduced from 2.1% to 1.0%.