I wrote a simple but fast image processing library to facilitate investigations in image enhancement. The library is written in C# and runs under the .NET framework.

The generic library code, along with code for a simple GUI to feed images and see the results of various computations, is in this repo.

All of the computations below involve separable filter kernels and should be easily implementable at video speeds on embedded hardware.

Original snapshot of finger showing fingerprint.
Gradient computation showing tangents & normals.
Gaussian blur followed by Laplacian, then a specialized thresholding.
Strength of computed orientation field allows us to mask off areas of the image that are not of interest.
This stable image results from repeatedly feeding back the output image as input.

That the image is a fixed point of this iterated function system suggests that it might be useful to make contraction mappings based on the original input, as a completely different way to proceed.