Beruflich Dokumente
Kultur Dokumente
CONTENTS
TOPIC
ABSTRACT
LIST OF FIGURES
CHAPTER-1: INTRODUCTION
1.1 Face Recognition Technology
1.2 Working of Face Recognition Technology
1.3 Introduction to Neural Networks
1.4 Use of Neural Networks
PAGE NO.
1-5
1
2
3
4
6-7
6
7
8-12
8
8
9
13-15
13
13
14
15
15
16
REFERENCES
18
LIST OF FIGURES
Figure 1.1 : Points of Recognition
Figure 2.1 : Components of Neuron
Figure 2.2 : The Synapses
Figure 2.3 : The Neuron Model
Figure 3.1 : A simple Neuron
Figure 3.2 : Basic Structure of Artificial Neural Network
Figure 3.3 : Network for Pattern Recognition
DEPARTMENT OF ECE
PAGE NO.
2
6
7
7
8
9
10
Page 1
10
11
12
12
13
14
CHAPTER-1
INTRODUCTION
1.1 Face Recognition Technology
Face recognition technology is the least intrusive and fastest biometric technology. It works
with the most obvious individual identifier the human face.
Instead of requiring people to place their hand on a reader(a process not acceptable in
some cultures as well as being a source of illness transfer) or precisely position their eye in
front of a scanner, face recognition systems unobtrusively take pictures of people's faces as
they enter a defined area. There is no intrusion or delay, and in most cases the subjects are
entirely unaware of the process. They do not feel "under surveillance" or that their privacy
has been invaded.
Facial recognition analyzes the characteristics of a person's face images input through
a digital video camera. It measures the overall facial structure, including distances between
eyes, nose, mouth, and jaw edges. These measurements are retained in a database and used as
DEPARTMENT OF ECE
Page 2
a comparison when a user stands before the camera. This biometric has been widely, and
perhaps wildly, touted as a fantastic system for recognizing potential threats (whether
terrorist, scam artist, or known criminal) but so far has not seen wide acceptance in high-level
usage. It is projected that biometric facial recognition technology will soon overtake
fingerprint biometrics as the most popular form of user authentication.
Every face has numerous, distinguishable landmarks, the different peaks and valleys that
make up facial features. Each human face has approximately 80 nodal points. Some of these
measured by the Facial Recognition Technology are:
DEPARTMENT OF ECE
Page 3
Corporations
Cash points
Stadiums
Public transportation
Financial institutions
Government offices
DEPARTMENT OF ECE
Page 4
DEPARTMENT OF ECE
Page 5
Page 6
6. There are several neural network models available to choose from in a particular
problem.
7. Due to increased accuracy, results in cost saving.
8. Neural networks can be retained using additional input variables and number of
individuals. Once trained they can be called on to predict in a new patient.
CHAPTER-2
Human and Artificial Neurons - Investigating the similarities
2.1 How the Human Brain Learns?
Much is still unknown about how the brain trains itself to process information, so
theories abound. In the human brain, a typical neuron collects signals from others through a
host of fine structures called dendrites. The neuron sends out spikes of electrical activity
through a long, thin stand known as an axon, which splits into thousands of branches. At the
end of each branch, a structure called synapse converts the activity from the axon into
electrical effects that inhibit or excite activity from the axon into electrical effects that inhibit
or excite activity in the connected neurons. When a neuron receives excitatory input that is
DEPARTMENT OF ECE
Page 7
sufficiently large compared with its inhibitory input, it sends a spike of electrical activity
down its axon. Learning occurs by changing the effectiveness of the synapses so that the
influence of one neuron on another changes.
Page 8
We conduct these neural networks by first trying to deduce the essential features of neurons
and their interconnections. We then typically program a computer to simulate these features.
However because our knowledge of neurons is incomplete and our computing power is
limited, our models are necessarily gross idealisations of real networks of neurons.
CHAPTER-3
AN ENGINEERING APPROACH
3.1 A simple neuron
An artificial neuron is a device with many inputs and one output. The neuron has two
modes of operation; the training mode and the using mode. In the training mode, the neuron
can be trained to fire (or not), for particular input patterns. In the using mode, when a taught
input pattern is detected at the input, its associated output becomes the current output. If the
input pattern does not belong in the taught list of input patterns, the firing rule is used to
determine whether to fire or not.
DEPARTMENT OF ECE
Page 9
DEPARTMENT OF ECE
Page 10
DEPARTMENT OF ECE
Page 11
X12:
X13:
OUT:
DEPARTMENT OF ECE
Page 12
X22:
X23:
OUT:
0/1
0/1
0/1
0/1
X22:
X23:
OUT:
From the tables it can be seen the following associations can be extracted:
DEPARTMENT OF ECE
Page 13
DEPARTMENT OF ECE
Page 14
CHAPTER 4
ARCHITECTURE OF NEURAL NETWORKS
4.1 Feed-forward networks
Feed-forward ANNs (figure 1) allow signals to travel one way only; from input to
output. There is no feedback (loops) i.e. the output of any layer does not affect that same
layer. Feed-forward ANNs tend to be straight forward networks that associate inputs with
outputs. They are extensively used in pattern recognition. This type of organisation is also
referred to as bottom-up or top-down.
Page 15
Page 16
case and is of more potential computational power than hierarchically structured multi-layer
organisations. In multi-layer networks, units are often numbered by layer, instead of
following a global numbering.
4.4 Advantages
When an element (Artificial neuron) of the neural network fails, it can continue
without any problem by their parallel nature.
A neural network learns and does not need to be reprogrammed.
It can be implemented in any application.
If there is plenty of data and the problem is poorly understood to derive an
approximate model, then neural network technology is a good choice.
There is no need to assume an underlying data distribution such as usually is done in
statistical modeling.
Neural networks are applicable to multivariate non-linear problems.
The transformations of the variables are automated in the computational process.
A neural network can perform tasks that a linear program cannot.
CHAPTER 5
CONCLUSION AND FUTURE SCOPE
DEPARTMENT OF ECE
Page 17
CONCLUSION
In this paper, we have proposed an intelligent face recognition system with the
following two desirable features: one-pass incremental learning and automatic generation of
training data. This system can be implemented with limited hardware constraints and is
suitable for the use in small devices where hardware capabilities are restricted such as home
security systems. A memory-based learning approach (accumulating and retraining all
previous data) may possibly obtain better recognition results but it is infeasible for large
datasets like image streams. From the online recognition experiments, we can conclude that
introducing the incremental learning function into the system is quite effective in the sense
that the system can enhance its recognition performance automatically without any
administrator intervention.
FUTURE SCOPE
Several open questions still remain in our face recognition system. First, the face
detection method introduced here is still rather immature in terms of the computation costs
and accuracy. Moreover, the robustness for image variations in rotations, size, illumination,
etc. must be improved.
Here, we evaluated the recognition performance only for small datasets of Japanese
people. From the aspect of security systems, such simple evaluations are useless. Hence, the
evaluation on the robustness for the larger datasets is necessary in practical use.
There are a number of directions for future work. The main limitation of the current
system is that it only detects upright faces looking at the camera. Separate versions of the
system could be trained for each head orientation, and the results could be combined using
arbitration methods similar to those presented here.
Preliminary work in this area indicates that detecting proles view of faces is more
difcult than detecting frontal views, because they have fewer stable features and because the
input window will contain more background pixels.
We have also applied the same algorithm for the detection of car tires and human eyes,
althoughmore work is needed.Even within the domain of detecting frontal views of faces,
more work remains.When an image sequence is available, temporal coherence can focus
attention on particular portions of the images.
DEPARTMENT OF ECE
Page 18
As a face moves about, its location in one frame is a strong predictor of its location
in next frame. Standard tracking methods, as well as expectation-based methods [2], can be
applied to focus the detectors attention. Other methods of improving system performance
include obtaining more positive examples for training, or applying more sophisticated image
preprocessing and normalization techniques.
REFERENCES
1. ELECTRONICS FOR YOU- Part 1 April 2001 & Part 2 May 2001
DEPARTMENT OF ECE
Page 19
DEPARTMENT OF ECE
Page 20