Ai Mini Project

You might also like

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 9

DEPARTMENT OF COMPUTER ENGINEERING

ARTIFICIAL INTELLIGENCE MINI PROJECT

MINI PROJECT TITLE


DIGIT RECOGNITION USING CNN

⚬Gauri Naik 21102A0015


GROUP MEMBERS : ⚬Rahil Shaikh 22102A2004
⚬Gayatri Vaishnav 22102A2006
GROUP ID : 11
⚬Sanika Kaginkar 22102A2008
INTRODUCTION
• Handwritten digit recognition is the ability of computers to recognize
human handwritten digits from various sources such as identifying
number plates of vehicles, handling bank cheques, digits entered in any
forms etc.
• The handwritten digits are not always of the same size or orientation and
position relative to the margins.
• It is tough task for the machine because handwritten digits are not easy
and perfect to understand easily, at the same time there is variety of hand
writings.
• Our system provides recognition of digits with accuracy up to 98 % to 99.9%.
ARCHITECTURE
ARCHITECTURE
Predicting categories for input data is called a classification task.
Classification tasks require an appropriate data representation for the labels
• Common representations of labels include one-hot encoding of categories
Prepare your data:
• It is useful to keep some data aside that the model never sees during training that you can use to evaluate the
model. This is called the validation set.
Build and run your model:
• Convolutional models have been shown to perform well on image tasks.
• Classification problems usually use categorical cross entropy for their loss functions.
• Monitor training to see whether the loss is going down and accuracy is going up.
Evaluate your model
• Decide on some way to evaluate your model once it's trained to see how well it is doing on the initial problem
you wanted to solve.
• Per class accuracy and confusion matrices can give you a finer breakdown of model performance than just overall
accuracy.
TECHNOLOGIES
1.Model Training:
⚬ Use PyTorch to define and train a convolutional neural network (CNN) model for digit recognition using the MNIST
dataset.
⚬ Save the trained model weights to disk for later use.
2.Flask Backend:
⚬ Set up a Flask web server to handle incoming requests.
⚬ Create routes for serving HTML pages and handling prediction requests.
⚬ Load the trained PyTorch model in the Flask backend.
⚬ Use NumPy and Pillow to preprocess drawn digit before passing them to the model for prediction.
⚬ Return the predicted digit as a response to the frontend.
3.HTML/CSS/JavaScript Frontend:
⚬ Design an HTML webpage with an input field for drawing digit.
⚬ Use CSS to style the webpage and make it visually appealing.
⚬ Implement JavaScript code to handle user interactions, such as drawn digit.
⚬ Display the predicted digit on the webpage.
SCREENSHOTS
FUTURE SCOPE

Further scope involves improving this model to recognize alphabets/character


strings by training it on a suitable dataset so that the usability of this system can be
extended to other domains (character recognition) as well.
REFERENCES
[1] “Handwriting recognition”: https : //en.wikipedia.org/wiki/Handwritingrecognition
[2] “What can a digit recognizer be used for?”: https : //www.quora.com/What can a digit recognizer be used for
[3] ”Handwritten Digit Recognition using Machine Learning Algorithms”, S M Shamim, Mohammad Badrul Alam Miah,
Angona Sarker, Masud Rana & Abdullah Al Jobair.
[4] ”Handwritten Digit Recognition Using Deep Learning”, Anuj Dutt and Aashi Dutt.
[5] ”Handwritten recognition using SVM, KNN, and Neural networks”, Norhidayu binti Abdul Hamid, Nilam Nur Binti
Amir Sharif.
[6] ”Recognition of Handwritten Digit using Convolutional Neural Network in Python with Tensorflow and Comparison of
Performance for Various Hidden Layers”, Fathma Siddique, Shadman Sakib, Md. Abu Bakr Siddique.
[7]. M. Wu and Z. Zhang, Handwritten Digit Classification using the MNIST Dataset, 2010.
[8]. A. Dutta and A. Dutta, Handwritten digit recognition using deep learning, International Journal of Advanced Research in
Computer Engineering & Technology (IJARCET), vol. 6, no. 7, July 2017.
[9]. Al Maadeed, Somaya, and Abdelaali Hassaine,Automatic prediction of age, gender, and nationality in offline
handwriting. EURASIP Journal on Image and Video Processing, no. 1 2014.
[10]. Gaurav Jain, Jason Ko, Handwritten DigitsRecognition, Project Report, University of Toronto, 11/21/2008.
THANK YOU

You might also like