Paimprint Recognition Using WaveletDecomposition and 2D PrincipalComponent Analysis
ErhuZhang JiwenLu, Xiaobing Kang · 2006
In thispaper,a novelmethodusingwavelet decomposition and2DPrincipal component analysis (2DPCA) forpalmprint recognition ispresented. Firstly, 2D wavelet transform isadoptedtoobtaindifferent levelofwavelet coefficients oftheoriginal palmprint image; secondly 2DPCA is appliedon the low-frequency thatcontains most discrimination information oftheoriginal palmprint image. One criterion thatnotallPCs areuseful forpalmprint recognition isdemonstrated andaruleforselecting 2DPCsis proposed. Lastly, thisalgorithm istested on thePolyU palmprint imagedatabase andtheexperimental result is encouraging and achieves comparatively highrecognition accuracy andmorecomputationally efficient thanusing other feature extraction techniques suchasprincipal component analysis andindependent component analysis.