Multispectral palmprint recognition using wavelet-based image fusion

Palmprint is widely used in personal identification for an accurate and robust recognition. To improve the existing palmprint systems, the proposed system, which is the first on-line multispectral palmprint recognition system ever designed before, uses multispectral capture device to sense images under different illumination, including red, green, blue and infrared. We adopt competitive coding scheme as matching algorithm, which performs well in on-line palmprint recognition. Wavelet-based image fusion method is used as data-level fusion strategy in our scheme. Fused verifications show better effort on motion blurred source images than single channel. Experimental results of fusion images are also useful references for future work on multispectral palmprint recognition.