Fingerprint Recognition Using Artificial Neural Networks
Fingerprint Recognition Using Artificial Neural Networks
Level: Level 4
Machine intelligence is the study of the principles, foundations, and designs of adaptive systems
that have the ability to learn, predict, optimise, and make decisions to accomplish goals through
interaction with uncertain environments.
In this paper we use a Multi-layer perceptron neural network with learning algorithm retro
propagation errors, for application in fingerprint recognition.
Traditionally, fingerprint recognition systems are based on the detection of individual fingerprint
traits that are measurable. These features are captured by a system that automatically performs
the recognition of the person and must meet the following conditions: universality (everyone
should have that feature), uniqueness (must be unique for each individual), stay (invariable
through time) and quantifiability (must be measurable in a quantitative manner).
The use of fingerprint biometric identification techniques is widely accepted because it is highly
reliable, easy to use, prevents against attacks, generally accepted, and reliable, it can be
authenticated, is cost efficient and standardized
Neural network
Neural network enables solutions to be found to problems where algorithmic methods are
too computationally intensive or do not exist
The problem of feature extraction and classification seems to be a suitable application for
neural nets.
They offer significant speed advantages over conventional techniques
Features of fingerprints
Fingerprints are imprints formed by friction ridges of the skin in fingers and thumbs.
Their pattern is permanent and unchangeable on each finger during all the life;
They are individual
They have long been used for identification
The pattern recognition system
Selection of good feature is a crucial step in the process since the next stage sees only
these features and acts upon them.
150 different minutiae type have been identified but in practice only ridge ending and
ridge bifurcation are used.
Good features are those satisfying two requirements:
Small interclass invariance (i.e. slightly different shapes with similar general
characteristics should have numerically close values)
Large interclass separation (i.e. features from different classes should be quite different
numerically)
Neural Networks Applied to the Detection of Fingerprints
A multilayer perceptron network of three layers is trained to detect the minutiae in the thinned
part image of size 128x128
The first layer has four units associated with the components of the input vector
The hidden layer has five units
The output layer has one unit corresponding to the number of the classes
The network is trained to output ‘1’ when the input window is centered on the feature to be
located and it outputs ‘0’ if minutiae are not present.
The network is trained by using the backpropagation learning technique and the weight
change is updated according to:
𝜕𝐸
∆𝑤(𝑛) = −𝜂 + 𝛼𝑤(𝑛 − 1)
𝜕𝑤
E is the energy function which is defined as the sum of the square difference between
the desired output response and the actual output response of each training example
The trained network is then used to analyse the complete image by raster scanning the fingerprint
via window of size 3x3
In order to prevent the falsely reported features and select “significant” minutiae, two more rules
are added to the system to guarantee perfect ridge forks are detected while excluding all other
features:
At those potential minutiae feature points we examine them by increasing the window
size to 5x5
If two or more minutiae are too close together, we ignore all of them