You are on page 1of 24

NEURO-FUZZY CLASSIFIERS

Project Guide:Ms. Apoorvi Sood


Assistant Professor

By:Monika Chhikara
12CSP032
M.TECH CSE-A

Introduction to soft computing


Fuzzy logic
Neural Network
Neuro fuzzy approach
Problem Statement
Combining Neural networks and fuzzy logic for
Pattern Classification
Review of papers
References

Soft Computing (SC): the symbiotic use of many


emerging problem-solving disciplines.
According to Prof. Zadeh:
"...in contrast to traditional hard computing, soft computing
exploits the tolerance for imprecision, uncertainty, and
partial truth to achieve tractability, robustness, low
solution-cost, and better rapport with reality

Constituents of Soft computing:


Fuzzy systems => imprecision
Neural networks => learning
Probabilistic reasoning => uncertainty
Evolutionary computing => optimization

Introduced by Lofti Zadeh(1965)


Fuzzy logic is used to define fuzziness
It is a problem solving methodology
Fuzzy logic was introduced to design
systems that can demonstrate human
like reasoning

Neural network: information processing


paradigm inspired by biological nervous
systems, such as our brain
Structure: large number of highly
interconnected processing elements (neurons)
working together
Like people, they learn from experience (by
example)

Why not make a system which learns linguistic rules from


input output data.

A neural network can learn from data.


But we cannot extract linguistic (or other easily
interpretable) rules from a trained network.

Can we combine these to paradigms?

YES!!

Neural networks can learn from data and feedback


Viewed as black box approach (know what the box
does but not how it is done conceptually!)
Unlike neural networks, fuzzy logic does not come
with a learning algorithm
Fuzzy rule-based models are easy to comprehend
(uses linguistic terms and the structure of if-then
rules)
Since neural networks can learn, it is natural to
combine the two technologies.

Problem statement ???


To develop an interactive fuzzy model of
HCR(handwritten character recognition) for
Devanagari Script.

Combining Neural networks and


fuzzy logic for Pattern
Classification

Optical Character Recognition:


Optical character recognition, usually abbreviated to OCR, is
the mechanical or electronic conversion of scanned images of
handwritten, typewritten or printed text into machine-encoded text.

Various steps for Character recognition


are:
1.
2.
3.
4.

PREPROCESSING
SEGMENTATION
CLASSIFICATION OF CHARACTER
FEATURE EXTRACTION

Preprocessing:
->The scanned image is first converted into binary
image containing 0 and 1 pixels only.
->Techniques like thinning , slant correction and
smoothing are applied

Segmentation:
It is done to make separation between the
individual characters of an image.
Each separated character generates a sub
image.

Classification Of Characters Of
Devanagari Script:
Devanagari Script can be classified in three major
categories:
Middle bar
End bar
Without bar

Feature Extraction:
The Box approach is used
This approach requires the spatial division of the
character image.
Each character image is divided into 24 boxes so
that the portions of a character will be in some of
these boxes.
All boxes are considered for analysis in a
sequential order

Architecture:

Literature survey

1.Neuro Fuzzy off-line Recognition of


Handwritten Sentences
To present a system for off-line recognition of
handwritten sentences using neural and fuzzy
logic technique.
System consists of 4 subsystems:
Preprocessing
Applying neural and fuzzy techniques
Lexical and grammatical analysis
Using words recognised with high confidence
Result
Around 98.7% context recognition.

2.Optical Character Recognition Using


Automatically Generated Fuzzy Classifiers
Problem :
To present a solution for the semi-automatic design of a Fuzzy classifier
for letters and digits to be applied on the automatic recognition of cars
license plates on unstructured conditions

Methodology:
GOAL: To create a system that could identify single font characters based on
their images captured by low cost digital image cameras
Useful in:
applications from automatic bank check ,numbering to cars license plate
recognition

Result:
98% for letter classifier
97% for digit classifier

3. Fuzzy Model based recognition of


handwritten Hindi characters
Problem :
To generate the recognition of handwritten Hindi Characters based on
the modified exponential membership function fitted to the fuzzy sets
derived from features consisting of normalized distances obtained using
the Box approach

Methodology:
A Reuse Policy that provides guidance from the past policies is utilized to
improve the reinforcement learning. This relies on the past errors exploiting
the past policies.
There is a 25-fold improvement in training. Experimentation is carried out on
a database of 4750 samples.

Result:
The overall recognition rate is found to be 90.65%

4. Interactive Fuzzy Model Based


Recognition of Handwritten Characters
Problem :
To generate an interactive fuzzy model based recognition for handwritten
characters.

Methodology:
In this fuzzy model based character recognition, the features extracted from a
character are used to form the fuzzy sets.
Assuming these fuzzy sets have the overlapping/interactive information, this
paper develops an Interactive Fuzzy Model by devising new fuzzy rules
for linking them.
The input fuzzy sets (formed from features) in the premise part of the
new fuzzy rule are represented by the modified exponential membership
function

Result:
99.4% for english numerals
95.8% for devanagari numerals

Structure of ocr system:

References:
M. Hanmandlu and O.V.R. Murthy Fuzzy Model Based
Recognition of Handwritten Hindi Characters, Proc. IEEE 2007.
Jose Manuel Fonseca, Nuno Miguel Rodrigues ,Andre Damas Mora
and Rita Almeida Ribeiro Optical Character Recognition Using
Automatically Generated Fuzzy Classifiers 2011 Eighth
International Conference on Fuzzy Systems and Knowledge
Discovery (FSKD)
M. Hanmandlu and O.V.R. Murthy Fuzzy Model Based
Recognition of Handwritten Hindi Characters, Proc. IEEE 2007.
Beatrice Lazzerini ,Francesco Marcelloni and Leonard0 Maria
ReyneriNeuro-fuzzy Off-line Recognition of Handwritten
Sentences Fourth Internntional Conference knowledge-Based
Intelligent Engineering Systems,Aug- Sept 2000

You might also like