Scribd
Upload
12 views37 pages

BATCH 6 For Presentation

The document presents a project on handwriting recognition using machine learning, specifically through a multilayer perceptron neural network. It discusses the limitations of existing OCR systems and proposes a new system utilizing convolutional neural networks for better accuracy and efficiency in character recognition. The project outlines system requirements, methodologies, implementation steps, and future enhancements for improving handwritten character recognition capabilities.

Uploaded by

manikumarande57
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
12 views37 pages

BATCH 6 For Presentation

The document presents a project on handwriting recognition using machine learning, specifically through a multilayer perceptron neural network. It discusses the limitations of existing OCR systems and proposes a new system utilizing convolutional neural networks for better accuracy and efficiency in character recognition. The project outlines system requirements, methodologies, implementation steps, and future enhancements for improving handwritten character recognition capabilities.

Uploaded by

manikumarande57
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 37

CHARACTER RECOGNITION USING MACHINE LEARNING

BY
A.Prathyusha -19KA1A0532
P. Bhavya -19KA1A0507
T. Pushpa -19KA1A0538
M .Ramya -19KA1A0503
T. Sreeja -19KA1A0530
M. Premalatha -19KA1A0555

Under the esteemed guidance of


Mrs.K.R.Lavanya,M.Tech.,
Assistant Professor(Adhoc)

Department of Computer Science and Engineering


Jawaharlal Nehru Technological University Anantapur College
Of Engineering, Kalikiri
CONTENTS
 INTRODUCTION
 ABSTRACT
 EXISTING SYSTEM
 PROPOSED SYSTEM
 SYSTEM REQUIREMENTS
 METHODS & STRATEGIES
 DESIGNS
 ARCHITECTURE OF THE MODEL
 RESULTS
 CONCLUSION
 FUTURE SCOPE
 REFERANCES
INTRODUCTION

 Handwriting Recognition is an ability of a computer to receive and interpret intelligible


handwritten input from sources .
 This project is intended to build a software that understands Hand-writing.
 The user writes with stylus or mouse in his own handwriting and it is recognized by the
system.
 This we are going to achieve by modeling a neural network that will
have to be trained over a dataset containing images of alphabets.
ABSTRACT

 The aim of this work is to review existing methods for the handwritten character
recognition problem using machine learning algorithms.
 The recognition model we have chosen is a multilayer perceptron, a feedforward
artificial neural network.
 The application was then tested on a training set consisting of digits with the ability to
learn alphabetical or different characters.
 It has also proved powerful in OCR and ICR systems that could be seen as a further
extension of this work.
EXISTING SYSTEM

 In the running world there is a growing demand for the users to convert the printed
documents in to electronic documents for maintaining the security of their data.
 The basic OCR system was invented to convert the data available on papers in to
computer process able documents, So that the documents can be editable and reusable.
 The existing system of OCR on a grid infrastructure is just OCR without grid
functionality.
 The existing system focuses more on background connectivity rather than text region.
i. Less efficient
ii. Time consuming
PROPOSED SYSTEM

 The proposed system is what we call grid infrastructure which eliminates the problem
of heterogeneous character recognition.
 In latest system we Neural Networks for image pre-processing and feature extraction.
 Convolutional Neural Networks that are used to extract the features of the images
using several layers of filters.
 The selection of the most feasible techniques and methods is highly dependent on the
available data.
Advantages of Proposed System

 The proposed system is more effective and accurate than other existing
systems.
 Comparison of various methodologies using different datasets and
characteristics.
 Tested with real time data.
System Requirements

 Software Requirements

 Operating System : Windows 10


 Coding Language : Python 3.8
 Tool : Jupyter Notebook
 Front End : GUI , Flutter
 Backend :Keras , OpenCV , TensorFlow
Hardware Requirements

 Processor : intel core i5


 Hard Disk : 250 GB
 Monitor : 15” LED
 Input Devices : Keyboard, Mouse
 RAM : 4 GB
CHARACTER RECOGNITION
Handwriting Recognition methods can be broadly classified into the below two types
METHODS & STRATEGIES

 Preprocessing of the dataset


 Segmentation
 Feature extraction
 Classification & Recognition
 Validation
Pre-processing of the data set

 Data Preprocessing is a technique that is used to convert the raw data into a clean data
set.

 Deletes all the duplicates data.


Segmentation

An image of the
sequence of characters
is decomposed into
sub images of
individual characters
Feature Extraction
The Feature Extraction technique explained was tested using a neural network which
was trained with features vectors obtained from the system proposed.
What is CNN?

 CNN stands for Convolutional Neural Networks that are used to


extract the features of the Images using several layers of filters.
Flow chart
USE CASE DIAGRAM
SEQUENCE DIAGRAM
Why convolutions?

 Convolutions are performed for various reasons

a. Convolutions provide better feature extraction


b. They save a lot of computation compared to ANN
c. Less number of parameters are created
IMPLEMENTATION
 LOADING DATA SET
i. A_Z Handwritten data set from Kaggle
ii. MNIST dataset from Sklearn
 Plotting the number of alphabets in the dataset
Split data into images and their labels:
# Split data the X - Our data , and y - the prdict label
X = data. Drop('0',axis = 1)
y = data['0’]

Reshaping the data in the csv file so that it can be


displayed as an image
• In the above segment, we are splitting the data into training &
testing dataset using train_test_split().
• Also, we are reshaping the train & test image data so that they can
be displayed as an image, as initially in the CSV file they were
present as 784 columns of pixel data. So we convert it to 28×28
pixels.
Shuffling the data
 The shuffling is done using the shuffle() function so that we can
display some random images.
 We then create 9 plots in 3×3 shape & display the thresholded
images of 9 alphabets.
Data Reshaping
 Reshaping the training & test dataset so that it can be put in
the model
 New shape of train data: (297960, 28, 28, 1)
 New shape of train data: (74490, 28, 28, 1)

Compiling & Fitting Model


Getting the Train & Validation Accuracies &
Losses
Doing Some Predictions on Test Data
 Here we are creating 9 subplots of (3,3) shape & visualize some of the test
dataset alphabets along with their predictions, that are made using
the model.predict() function for text recognition.
Doing Prediction on External Image
GUI IMPLEMENTATION
 Live Cropped Character Recognition Application Window
DRAW CHARACTERS
CROP CHARACTER & PREDICT
CONCLUSIONS
 The Handwritten character recognition using convolutional
neural network has proved to be a fairly good efficiency.
 It works better than any other algorithm , including artificial
neural networks.
 Improve the system to identify any other characters.
FUTURE SCOPE

 In future work, we train the model to read a whole


handwritten or non-handwritten document and convert
the printed text to different languages.
 We also incorporate a personalized scanned file storage
set-up for each registered user. The model can be trained
further for specific purposes.
REFERENCES

 [1] H. Zeng,(2020)An Off-line Handwriting Recognition Employing Tensorflow.


International Conference on Big Data, Artificial Intelligence and Internet of Things
Engineering (ICBAIE).
 [2] A. Beltran and S. Mendoza, “Efficient algorithm for real-time handwritten
character recognition in mobile devices “,2011, 8th International Conference on
Electrical Engineering, Computing Science and Automatic Control, 2011, pp. 1-6, doi:
10.1109/ICEEE.2011.6106583.
 H. Du, P. Li, H. Zhou, W. Gong, G. Luo and P. Yang, "WordRecorder: Accurate
Acoustic-based Handwriting Recognition Using Deep Learning,"IEEE INFOCOM
2018 - IEEE Conference on Computer Communications, 2018, pp. 1448-1456, doi:
10.1109/INFOCOM.2018.8486285.
THANK YOU

You might also like