Fingerprint image denoising using morphological amoebas

This paper deals with noisy problem in the fingerprint images. With the help of pilot images from Canny edge detection, the method of morphological amoebas can simultaneously reduces noises and keeps useful details by filtering images with non-fixed kernels, outperforms classical morphological operations and other filters with fixed, space-invariant structuring element. What´s more, even the approach needs computing the structuring elements for every point in the image, the run-time speed of this algorithm is still faster than some nonlinear algorithms. Finally, a practice with satisfactory result proves that this method do work effectively.