Experimental approach on thresholding using reverse biorthogonal wavelet decomposition for eye image

This study focus on compression in wavelet decomposition for security in biometric data. The objectives of this research are two folds: a) to investigate whether compressed human eye image differ with the original eye and b) to obtain the compression ratio values using proposed methods. The experiments have been conducted to explore the application of sparsity-norm balance and sparsity-norm balance square root techniques in wavelet decomposition. The eye image with [320x280] dimension is used through the wavelet 2D tool of Matlab. The results showed that, the percentage of coefficients before compression energy was 99.65% and number of zeros were 97.99%. However, the percentage of energy was 99.97%, increased while the number of zeros was same after compression. Based on our findings, the impact of the compression produces different ratio and with minimal lost after the compression. The future work should imply in artificial intelligent area for protecting biometric data.