作者: Deepti Tamrakar , Pritee Khanna
DOI: 10.1007/S11042-015-2541-5
关键词:
摘要: Rotation and noise invariant feature extraction is a challenge in palmprint recognition. This work presents novel RDF descriptor based on Radon, Dual tree complex wavelet, Fourier transforms. Combined properties of these transforms help to explore efficiency robustness for identification. Radon transform can capture directional features the robust additive white Gaussian also. It converts rotation into translation. 1D wavelet (DTCWT) applied coefficients angle direction removes translation due rotation. The magnitude 2D performed resultant helps extract illumination features. performance proposed evaluated noisy rotated palmprints upto 10?. Trained with normal only, system gives good results palmprints. Experiments are PolyU 2D, CASIA, IIITDMJ databases. Theoretical foundations experimental show against