Fingerprint Biometric Detection

Overview

This C# project explores fingerprint matching by representing fingerprint images as binary and ASCII strings. The image-processing code separates ridge and valley pixels using a threshold and prepares image-derived patterns for string matching. The Windows Forms application loads a collection of fingerprint images from a MySQL-backed database and converts those images concurrently before presenting the data to the interface.

Pattern matching

The matching code searches an image's ASCII representation for patterns from the stored images. Knuth-Morris-Pratt preprocesses a pattern with a border-function table so it can reuse prefix matches after a mismatch. Boyer-Moore compares from the end of the pattern and uses a last-occurrence table to skip ahead. Matches contribute weighted points, and the implementation uses Levenshtein distance to find a closest candidate when none of the patterns match exactly.

if (KMP.KMPStringMatching(imageAscii, patterns[i]))
{
    matches += CalculateWeight(i + 1);
}

This is an exploration of adapting text-search algorithms to image-derived data, not a claim of production-grade biometric verification. The key learning goal was to connect preprocessing, string representation, exact pattern matching, and approximate comparison in one application.