Comparative Analysis of Coiflet and Daubechies Wavelets in Fingerprint Image Compression.

The choice of wavelet bases for image signal decomposition and energy de-correlation, as well as, the choice of threshold strategy to achieve sparse representation for image signal are very crucial in the overall performance of a wavelet-based image compression system. In this comparative study, orthogonal Daubechies and Coiflet wavelet bases were applied at different orders and decomposition levels to decompose fingerprint image into approximation and detail components. Level-dependent threshold was applied at three different decomposition levels using Birge-Massart strategy to achieve a sparse representation of the image and the performance of Daubechies and Coiflet wavelets for compression algorithm were evaluated. The performances of the investigated methods were evaluated on the basis of percentage number of zero, NZ (%) and retained energy, RE (%). At decomposition level three, 8-bit source fingerprint image was completely decomposed and the RE (%) values for representation based on Coiflet order 4 ranged from 99.32% to 99.69% as opposed to the values for Daubechies order 4 which ranged from 98.45% to 98.91%. These results revealed that Coiflet wavelet bases performed better than Daubechies wavelet bases in terms of the capability for image energy decorrelation and sparse representation.