Filtering Algorithm of the False Features for Fingerprint Images Based on Wavelet Transform

According to the characteristics of fingerprint images, combined with the human visual properties, this paper presents a de-noising algorithm of fingerprint image processing to decrease the influence of noises. Based on the different characteristics between the noise points and image features, the algorithm decreased the noises of the source images and filtered the corresponding noise coefficients after the wavelet decomposition. As for the coefficient matrices following the wavelet decomposition, after calculating all local gradients of the coefficients, we established the algorithm by using the local gradients of the source images as the judgment basis, and chose the highest Gradient coefficient in the different directions of the source images as the final fusion coefficient. The experimental results indicate that the algorithm can promote the SNR of the source images, protect the details of the images and improve the visual effects.