Marcos Rios

David Garcia

Mark Valtierra

 

Objective

The objective of the Automated Fingerprint Identification System (AFIS) is to attempt to classify fingerprint patterns into one of five major classes.

AFIS Goals Reached

  1. The project was able to capture a person's fingerprint pattern via a scanner and store it in some sort of database.

  2. The project was able to map the image of the fingerprint to a particular class.

  3. The project has the options of zooming out and in on the fingerprint image.

  4. The project has the option of printing a fingerprint image onto a paper.

Method of Completion

  1. Filtering of Fingerprint Image: It was done to a rolled inked fingerprint image acquired by means of a flatbed scanner.  This inked fingerprint image had to be converted to a gray scale image format from an RGB image representation so it could be filtered.

  2. Feature Extraction of Fingerprint Image: In order for feature extraction to be effective, the image had to be turned from the gray scale image into a binary image.  The feature extraction was responsible for reproducing the feature vector.

  3. Neural Network Algorithm: Its purpose was to classify a set of fingerprint images to a particular fingerprint class via the Learning Vector Quantization (LVQ) algorithm.

Issues

  • This project did not meet the expectations specified in the original proprosal.  Originally it was hoped that the AFIS would be able to both classify and match a fingerprint to his/her respective owner.  However, this proved to be a formidable task and it was beyond our scope at this present writing.

  • Although, the AFIS didn't meet the requirements stated in the proposal, some positive results were generated.  The Filtering and Preprocessing sub processes yielded favorable results.  The Filtering module was able to adequately take a gray scale fingerprint image and convert it into a binarized fingerprint image.  The Preprocessing performed as it was intended to do so; it did generate the feature vector each time for each fingerprint pattern with relative ease.

Block Diagram for Fingerprint

Conclusion    

This Particular AFIS was able:

  • To acquire a rolled inked fingerprint image.

  • To extract feature information from the binary image to create a feature vector.

  • To print a filtered fingerprint image.

  • To zoom in and zoom out on an image.

  • To classify a set of fingerprint images to a particular fingerprint class.