Fingerprint Minutiae Extraction & Matching

A Java-based image processing library for biometric fingerprint analysis. This project implements the full pipeline from raw boolean images to feature-based matching without external CV libraries.

🛠 Features

• Iterative Thinning: Implements a two-step morphological thinning algorithm to extract the 1-pixel wide skeleton of ridges.
• Minutiae Extraction: Detects ridge endings and bifurcations using transition counts (Crossing Number algorithm).
• Geometric Analysis: - Orientation Estimation: Computes ridge flow angles using linear regression over local pixel clusters.
  • Spatial Transformations: Supports rotation and translation matrices to align disparate fingerprint captures.
• Feature Matching: A threshold-based matching engine that compares minutiae sets across different coordinate systems.

📐 Mathematical Approach

The project relies on several key mathematical concepts:

1. Linear Regression: Used in computeSlope to determine the best-fit line for ridge orientation.
2. Euclidean Distance: For calculating proximity between feature points.
3. Zhang-Suen Morphological Steps: To preserve topological properties during image thinning.

🚀 Usage

The core logic is contained within Fingerprint.java.

// Extract features from a thinned image
List<int[]> minutiae = Fingerprint.extract(skeletonImage);

// Compare two sets of minutiae
boolean isMatch = Fingerprint.match(minutiaeSetA, minutiaeSetB);