You are on page 1of 29

“If the brain were so simple that

we could understand it then we’d


be so simple that we couldn’t”
DEVELOPMENT OF
FLANN BASED
ADAPTIVE DIGITAL
IMAGE FILTER
TOPICS TO BE DISCUSSED
1)IMAGE AND IMAGE PROCESSING

2)FILTERS AND TYPE OF FILTERS

3)DIFFERENT TECHNIQUES USED


TO DESIGN A FILTER
4)BEST TECHNIQUE
What is Image?
The term image refer to a two
dimensional light intensity function
f(x,y), where x and y denote spatial co-
ordinates and the value of f at any point
(x,y) is proportional to the brightness (or
gray level) of the image at that point.
What is Image
processing
Image processing is any form of
information processing for which
the input is an image, such as
photographs or frames of video;
the output is not necessarily an
image, but can be for instance a
set of features of the image.
Image Processing
Techniques
Image Acquisition
Image Enhancement
Image Restoration
Image Compression
Image Segmentation
What is Image
Restoration
It is an Image processing technique for
getting an original image from a noisy
image.
It is of two 2 Types:

1.Image Denoising

2. Image Deblurring
Different filter used for
Image Denoising

1.Linear Filters

2.Non-Linear filter
Drawbacks of linear
filter
Linear filters generally have Linear
characteristics. But on online application the noise
added to the system is adaptive in nature. On that
case the linear filter won’t able to denoise.
Non linear filter

In order to restore the image from


adaptively degraded image, we are
generally using non linear filter.
Different Algorithms For
Designing Non Linear Filters
• Neural Network
• Fuzzy logic
• Ant colony
• Bacteria Technology
• Genetic algorithm

But we prefer Neural Network


INTRODUCTION TO
NEURAL NETWORK
 It is an interconnected group of artificial neuron
that use a mathematical model or computational
model for information processing based on a
connectionist approach to computation.

 Its main aim is to mimic the human ability to adapt to


changing circumstances and current environment.
Basic Concepts
• A Neural Network
Input 0 Input 1 ... Input n
generally maps a set of
inputs to a set of outputs

• Number of inputs/outputs Neural Network


is variable

• The Network itself is Output 0 Output 1 ... Output m

composed of an arbitrary
number of nodes with an
arbitrary topology
What can Neural Network do?
• Compute a known function
• Approximate an unknown function
• Pattern Recognition
• Signal Processing
NEURAL NETWORKS ARE CATEGORISED
MAINLY INTO THREE PARTS

1)SINGLE LAYER 2)MULTI LAYER


PERCEPTRON PERCEPTRON
3) FLANN - FUNCTIONAL LINK
ARTIFICIAL NEURAL NETWORK
Single Layer Perceptron
(9-1 network)
MULTI LAYER
9-4-1 Neural Network
FLANN
W

X d
1

X
2 e
y
-
-
-
-
X
9

LMS
DIFFERENT FUNCTIONAL
ENHANCING PATTERNS ARE :
Exponential expansion
Trigonometric Functional expansion
Power series expansion
Chebyshev expansion
WORKING WITH SINGLE LAYER
PERCEPTRON
WORKING WITH MULTI LAYER
PERCEPTRON
WORKING WITH FLANN

DISTORTED IMAGE
ORIGINAL IMAGE Image corrupted with RESTORED IMAGE
Gaussian Noise of mean USING FLANN
0 and variance 0.01
COMPARISION OF MLP &FLANN
Status of Neural Networks
• Most of the reported
applications are still in research
stage
• No formal proofs, but they seem
to have useful applications that
work
Conclusion
It can concluded that the performance of FLANN
is better than MLP for noise suppression from an
image. The FLANN structure may be used for
online image processing application due to its less
computational requirement and satisfactory
performance.
References
[5.1] P.J.Antsaklis,"Neural network in control system."IEEE,
Control.Syst.Mag.PP. 3-5. April.1990.
[5.2] S. Haykin. Neural Networks. Ottawa.ON.Canda Maxwell Macmillan.
1994.
[5.3] P.S.Sastry.G . Santharam and K.P. Unnikrishnan." Memoryneural
networks for identification and control of dynamical systems." IEEE
Trans.Neural Networks. vol. 5.pp. 306-319.Mar.1994.
[5.4] A.G. Parlos.K.T.Chong and A.F .Atiya."application of recurrent
multilayer perceptron in modeling of complex process dynamics."
IEEE Trans. Neural Networks .vol.5,pp255-266,Mar.1994.
[5.5] R Grino.G.Cembrano and C.Torres."Nonlinear system
identification using additive dynamic neural networks two on line
approaches."IEEE Trans Circuits SystemIvol47 150-165.Feb 2000.
[5.6] T.Poggio and F.Girosi."Networks for approximation and
learning."Proc. IEEE ,vol 78,pp.1481-1497,sep1990.
TUHIN GHOSE
&
KEERTI PRAKASH PARIJA

TRIDENT ACADEMY OF
TECHNOLOGY
BHUBANESWAR

You might also like