![Reference No. Paper Title Research methodologies Scope of paper
[1] Recognition of Handwritten
Digit using Convolutional
Neural Network in Python
with Tensor flow and
Comparison of Performance
for Various Hidden Layers
In this paper, To find the
performance of CNN hidden layers,
CNN as well as MNIST dataset
have used. The network is trained
using stochastic gradient descent
and the backpropagation algorithm.
Scope of this paper is to
adding feature like GUI to take
real time input.
[2] Analogizing Time
Complexity of KNN and
CNN in Recognizing
Handwritten Digits.
In this paper comparison of
accuracy of KNN as well as CNN
have done. In this paper MNIST
data set has used for providing
samples to the system.
Scope of this paper is to
choosing CNN because it
produces high accuracy then
KNN. And also adding GUI
for real time input.
Background](https://image.slidesharecdn.com/a-22ppt-210712084714/85/GUI-based-handwritten-digit-recognition-using-CNN-4-320.jpg)
![Reference No. Paper Title Research methodologies Scope of paper
[3] Study and Observation of the
Variations of Accuracies for
Handwritten Digits Recognition
with Various Hidden Layers and
Epochs using Neural Network
Algorithm
In this paper comparison of two
neural network have done. In this
paper comparison has done with
respect to hidden layers, batch
size and epochs and also
explained about complexity of
ANN.
Scope of paper is to planning
about using of convolutional
neural network instead of
artificial neural network
because CNN is less
complex as well as more
efficient then ANN.
[4] Handwritten Digit Recognition
using CNN
In this paper comparison of KNN
, SVM, RFC and CNN have done.
Scope of this paper planning
to use CNN and there also
use of tensor flow as
backend for creation of
model. Also make GUI
which gives handwritten
digit to system as input for
classification.
Background continued..](https://image.slidesharecdn.com/a-22ppt-210712084714/85/GUI-based-handwritten-digit-recognition-using-CNN-5-320.jpg)
![Reference
[1] Fathma Siddique and Shadman Sakib, “Recognition of Handwritten Digit using
Convolutional Neural Network in Python with Tensor flow and Comparison of
Performance for Various Hidden Layers” CSE, EEE, March 2019.
[2] Tanya Makkar and Yogesh Kumar, “Analogizing Time Complexity of KNN and
CNN in Recognizing Handwritten Digits”, ICIIP, March 2017.
[3] Md. Abu Bakr Siddique1 and Mohammad Mahmudur Rahman Khan, “Study and
Observation of the Variations of Accuracies for Handwritten Digits Recognition with
Various Hidden Layers and Epochs using Neural Network Algorithm”, EEE, ECE,
March 2018.](https://image.slidesharecdn.com/a-22ppt-210712084714/85/GUI-based-handwritten-digit-recognition-using-CNN-15-320.jpg)
![[4] Vijayalaxmi R Rudraswamimath and Bhavanishankar, “Handwritten Digit
Recognition using CNN”, ISSN, June 2019
[5] Kecheng Wang and Junwen Deng, “The Four Arithmetic Operations for
Handwritten Digit Recognition Based On Convolutional Neural Network”,
FIM, WU, February 2020.
Reference](https://image.slidesharecdn.com/a-22ppt-210712084714/85/GUI-based-handwritten-digit-recognition-using-CNN-16-320.jpg)
GUI based handwritten digit recognition using CNN
- 1. Mr. Durgesh R Pandey (BE E&TC A 67) Mr. Abhishek M Tiwari (BE E&TC A 73) Mr. Ashwani V. Gupta (BE E&TC A 26) Mrs. Purnima Chandrasekhar (Mentor)
- 2. Index
- 3. Introduction • There are lots of handwritten documentation work in daily life and still machine have difficulties to recognize that handwritten digits as well as characters. There are different styles to write a single digit because of that, machine can not recognize the digit. • So the goal of this project is to create a model which can recognize the digits as well as also to create GUI which is user friendly i.e. user can draw the digit on it and will get appropriate output.
- 4. Reference No. Paper Title Research methodologies Scope of paper [1] Recognition of Handwritten Digit using Convolutional Neural Network in Python with Tensor flow and Comparison of Performance for Various Hidden Layers In this paper, To find the performance of CNN hidden layers, CNN as well as MNIST dataset have used. The network is trained using stochastic gradient descent and the backpropagation algorithm. Scope of this paper is to adding feature like GUI to take real time input. [2] Analogizing Time Complexity of KNN and CNN in Recognizing Handwritten Digits. In this paper comparison of accuracy of KNN as well as CNN have done. In this paper MNIST data set has used for providing samples to the system. Scope of this paper is to choosing CNN because it produces high accuracy then KNN. And also adding GUI for real time input. Background
- 5. Reference No. Paper Title Research methodologies Scope of paper [3] Study and Observation of the Variations of Accuracies for Handwritten Digits Recognition with Various Hidden Layers and Epochs using Neural Network Algorithm In this paper comparison of two neural network have done. In this paper comparison has done with respect to hidden layers, batch size and epochs and also explained about complexity of ANN. Scope of paper is to planning about using of convolutional neural network instead of artificial neural network because CNN is less complex as well as more efficient then ANN. [4] Handwritten Digit Recognition using CNN In this paper comparison of KNN , SVM, RFC and CNN have done. Scope of this paper planning to use CNN and there also use of tensor flow as backend for creation of model. Also make GUI which gives handwritten digit to system as input for classification. Background continued..
- 6. • Block diagram Proposed Architecture
- 7. • Implementation of system divided into 2 parts: CNN model: It is the main part of the system, which is present in the backend of the system. By this system can classify the digit. GUI window: GUI window is the part of frontend of the system. By using this user can draw the digit on it, which is taken as input to the system which will further go for classification as well as recognition to CNN. Implementation / Testing continued..
- 8. Flow diagram Backend (Developer) Frontend (User) MNIST dataset Preprocessing Preprocessing Training and testing of CNN model Hidden Layer Fully Connected Layer Trained CNN Model Saving and loading into GUI Going for Classification Result with accuracy DRAW DIGIT HERE….
- 9. • Implementation of CNN model part is divided into 5 parts: i. Importing of libraries and load data set. ii. Preprocessing of data set. iii. Creation of model. iv. Training of model. v. Evaluation of model. Implementation / Testing continued..
- 10. Results of CNN
- 11. Confusion matrix
- 12. • Import model and libraries. • Load model • Preprocessing • Canvas (elements & grid) creation • Grabbing image • implementation GUI (GRAPHIC USER INTERFACE)
- 13. OUTPUT
- 14. • There are problem with machine to recognize the handwritten digit. So this system will provide a CNN model which can recognize digit and also provide GUI to make user friendly environment, Because of the CNN this system has become more reliable, less complex and easy to understand as well as GUI gives the window to user, by which user can draw the digit on it. • This project will help to reduce machine efforts to recognize handwritten digits. • This project is design to make learning process easy and efficient. Conclusion
- 15. Reference [1] Fathma Siddique and Shadman Sakib, “Recognition of Handwritten Digit using Convolutional Neural Network in Python with Tensor flow and Comparison of Performance for Various Hidden Layers” CSE, EEE, March 2019. [2] Tanya Makkar and Yogesh Kumar, “Analogizing Time Complexity of KNN and CNN in Recognizing Handwritten Digits”, ICIIP, March 2017. [3] Md. Abu Bakr Siddique1 and Mohammad Mahmudur Rahman Khan, “Study and Observation of the Variations of Accuracies for Handwritten Digits Recognition with Various Hidden Layers and Epochs using Neural Network Algorithm”, EEE, ECE, March 2018.
- 16. [4] Vijayalaxmi R Rudraswamimath and Bhavanishankar, “Handwritten Digit Recognition using CNN”, ISSN, June 2019 [5] Kecheng Wang and Junwen Deng, “The Four Arithmetic Operations for Handwritten Digit Recognition Based On Convolutional Neural Network”, FIM, WU, February 2020. Reference