Security

Multimodal Biometric Authentication System Based on High Level Feature Fusion Approach

Free registration required

Executive Summary

Multimodal biometric systems fuse the information from the several biometric modalities to obtain better verification/ identification performance. The practical deployment of biometric identification/ verification systems presents a number of challenges, among which the need in many applications for high levels of accuracy. In this paper, an effective fusion scheme is presented that combines information presented by multiple domain experts based on the rank-level fusion integration method. The developed multimodal biometric system possesses a number of unique qualities, starting from utilizing Fisher's Linear Discriminant methods for face matching, minutia features for fingerprint matching and local binary pattern features for iris matching and fused the information for effective recognition and authentication.

  • Format: PDF
  • Size: 317.69 KB