Fingerprint Recognition System is a type of Pattern Recognition System that is widely used in Bio-metrics. This post will discuss in detail about what is Fingerprint Recognition System, its architecture, how it works, applications & advantages and disadvantages.
What is Fingerprint Recognition System
Fingerprint Recognition System is a Bio-metric Recognition System that recognizes a person by his Fingerprint. This system is used for identification and verification purpose. The user whose impression is to be identified or verified; his/her Fingerprint is registered onto the Fingerprint Scanner by placing fingertip on the scanner. Fingerprint is the impression of Fingertip and the scanner produces digital representation of the same and then it is compared with the Database template and thus the user is identified and verified.
Fig. 1 – Introduction to Fingerprint Recognition System
Digital Payment Application uses Fingerprint Recognition System and Fig. 2 shows steps involved in the Payment Application. In the First stage, the User requests for online transfer to specific recipient. Local Device (Smart Phone) asks for a Bio-metric or Finger impression and once the impression matches with the stored image in the Database, transfer of amount takes place. The user gets a password less experience that saves his time in remembering password.
Fig. 2 – Representation of Digital Payment Transaction
Architecture of Fingerprint Recognition System
The components of this system are:
- Enrolment Module
- Processing Module
- Database Module
- Verification/Identification Module
This Module deals with registering the Users Fingerprint. Fingerprint Scanner scans the Finger impression and produces raw digital representation.
Processing stage of the system accepts the input from the scanner and processes it further. The Fingerprint undergoes Feature Extraction and generates Feature Vectors.
The Users templates are stored in the Database Module. The Feature Vector generated by the Processing Module is compared against one or more existing templates.
This Module interfaces with the application system and the User thus claims his/her identity.
Fig. 3 – Architecture of Fingerprint Recognition System
How does Fingerprint Recognition System Work
To understand the working principle of the system, Let us consider Fig. 4. It shows different processes involved in identifying a person using his Fingerprint. Fingertip contains Friction Ridges and Furrows which are unique & persistent which is very helpful in Biometric Technology.
First step involves acquisition of the Fingerprint using Scanner. User has to place his finger on the scanner which has a sensor. To capture a person’s Finger impression, Optical and Solid state sensors are mainly used.
Every persons Finger impression is unique and can be distinguished by Minutia, which are some abnormal points on the Ridges. Minutia extraction involves three stages.
- Pre- Processing stage
- Minutia Extraction stage
- Post Processing stage
This stage is sub divided into 3 stages.
- Image Enhancement
- Image Binarization
- Image Segmentation
Fig. 4 – Representation of Pre-Processing Stage
This stage involves Image Enhancement through Contextual filtering method such as Gabor Filtering. Finger impression acquired must be enhanced as the ridges may be cut or broken, due to excess Finger pressure on the sensor, adjacent Ridges may appear joined.
Once the image is Enhanced, it has to be binarized. Hence this step is referred as “Image Binarization”. This is done by comparing the thresholds which is a combination of variance and local intensity Mean.
Image Segmentation involves excluding the end points of the ridge lines at the boundary of the Finger impression to avoid the extraction of false or unwanted features.
Minutia Extraction stage
This stage is sub-divided into 2 stages. They are:
- Minutiae Marking
Thinning process reduces Ridge thickness to one pixel that allows Minutiae detection. Spurious imperfections may occur due to Thinning process and it is further processed to remove the imperfections.
Minutiae Marking is done using Crossing Number (CN) method i.e. it uses 3 x 3 Pixel Window to classify Ridge Pixel. Crossing Number (CN) Value is computed based on the Crossing Number properties as shown in the table (Fig. 5(a)) below:
Fig. 5 – (a) Crossing Number Properties (b) Sample Fingerprint showing Ridge Characteristics
Suppose a Ridge Pixel with CN of 0 will correspond to an isolated point and a CN of 1 corresponds to Ridge Ending Point.
Post Processing Stage
This stage involves two steps:
- Removal of False Minutiae
- Fingerprint Matching
Fig. 6 –Working Principle of Fingerprint Recognition System
Removal of False Minutiae
The Finger impression which contains islands, Delta, holes are discarded. Side Minutiae, hooks, overlaps, Minutiae in regions of poor image quality, minutiae that are too wide and too narrow pores are removed.
Fingerprint Matching is done using BOZORTH3 Algorithm. It helps in matching of the input Finger impression received from sensor with the stored Templates in Database and outputs the result.
Applications of Fingerprint Recognition System
The applications of Fingerprint Recognition System include:
- It is used in identification of specific users.
- It is widely used in Digital Payment applications.
- Fingerprint is mandatory for issuing of Passports.
- It is used in banking Sector.
- It is used in National Identification System.
- This is the most accepted system for Voting purposes.
- It is used in Physical Access Control Solutions.
- It is used in Security Systems.
Advantages of Fingerprint Recognition System
The advantages of Fingerprint Recognition System are:
- This system is economical.
- The system is very reliable and stable.
- Output is more accurate.
Disadvantages of Fingerprint Recognition System
The disadvantages of this system are:
- Implementation of the system is tough.
- Fingerprint scanners do not work when the Fingertip is dirty or wet.
- Though the system is stable, Fingerprints vary with age and environmental conditions.