Training Report

On

Digit Recognition
Submitted in the fulfillment of the
Requirement for the award of the degree of

BACHELORS OF COMPUTER APPLICATIONS

(Batch 2022-2025)

By

Chhavi
1322190

Under the supervision

Of
Ms. Rubika Walia

M.M. Institute of Computer Technology & Business Management

Maharishi Markandeshwar (Deemed to be University), Mullana (Ambala)
(NAAC accredited Grade ‘A++’ University)
DECLARATION
I hereby declare that the work which is being presented in
this report entitled “DIGIT RECOGNITION”, in partial
fulfillment of the requirement for the award of the degree of
BACHELOR OF COMPUTER APPLICATIONS submitted at M.M.
Institute of Computer Technology & Business Management,
Maharishi Markandeshwar (Deemed to be University),
Mullana, Ambala is an authentic work done by me during a
period from 3July2024 To 20August2024 under the Guidance
of DR. RUBIKA WALIA.

The work presented in this report has not been submitted by

me for the award of any other degree of this or any other
Institute/University.

Signature
Chhavi
1322190

This is to certify that the above statement made by the

candidate is correct to best of my knowledge and belief.

Date: Signature
Place: Dr. Rubika Walia
Associate proff.
COMPANY CERTIFICATE ON COMPANY’S LETTER HEAD
ACKNOWLEDGEMENT

I would like to express my sincere gratitude to everyone who

contributed to the successful completion of my project on
the "Digit Recognition using Python and MySQL."

Firstly, I would like to thank my Professor Dr Rubika Walia

for their invaluable guidance, support, and encouragement
throughout the course of this project. Their insights and
expertise were crucial in helping me overcome challenges
and refine my ideas.

I also extend my thanks to my classmates and friends, who

provided their feedback and suggestions, which greatly
improved the quality of my work.

Furthermore, I am grateful to my teacher and family for their

unwavering support and understanding, which gave me the
motivation to see this project through to completion.

Lastly, I would like to acknowledge the online resources and

communities that provided valuable information and
solutions, which were instrumental in resolving technical
issues encountered during the development of this project.

ABSTRACT
List of diagram with page number
1.logo
2.Data flow diagram
3.Er model
4.implementing
5.testing

List of table with page number

No table

List of Abreviations
AI
GPU
API

INDEX
1.Introduction
a) Organization profile
b) Introduction of project
c)Problem definition
d) Limitation of existing system
e) Objective of the project

2. System Analysis
a) Feasible study
b) Hardware and software requirements

3. Development Environment
a) Introduction of technology used

4. System Design
a) Modular description

5. Implementation and Testing

a) Implementation and screenshots
b) Testing

6. Conclusion

7. Bibliography

1.Introduction
 A) Organization profile

Excellence Technology is one of the most well-known and

respected software development companies in Chandigarh
and Mohali that is ISO 9001: 2015 certified . They create
websites, games, antivirus software, and mobile
applications. We offer the best industrial training available in
Mohali and Chandigarh. They work on software testing,
networking, Android, Java, PHP, and web design. There
offshore clients are from Australia, the United States,
France, Germany, Finland, Madagascar, Sudan, and
Germany. There offshore clients' real projects will provide
our students with an opportunity to receive their industrial
training. With a comprehensive range of courses in software,
networking, automation, digital marketing, MBA, and many
more, they provide reasonable prices for website design and
development at Excellence Technology. There modern
facilities include roomy classrooms. Additionally, we give the
best job-oriented courses in Chandigarh that are 100%
practical and provide stipends for real-world projects. Pupils
that meet the requirements can receive instruction in digital
marketing, web design, PHP, and android in Chandigarh,
along with an internship or training that pays. We guarantee
a 100% employment with a two-year course validity. There
main goal is to provide each new hire with a competitive
compensation package. They offer the best industrial
training available in Mohali and Chandigarh. They work on
software testing, networking, Android, Java, PHP, and web
design. There offshore clients are from Australia, the United
States, France, Germany, Finland, Madagascar, Sudan, and
Germany. There offshore clients' real projects will provide
our students with an opportunity to receive their industrial
training. They give reasonable prices for website design
services. & Development, offering a variety of courses in
they give our pupils access to the greatest learning
environment possible so they may advance their abilities.
Additionally, they give the best job-oriented courses in
Chandigarh that are 100% practical and provide stipends for
real-world projects. Pupils that meet the requirements can
receive instruction in digital marketing, web design, PHP,
and android in Chandigarh, along with an internship or
training that pays. They guarantee a 100% employment
with a two-year course validity. Our main goal is to provide
each new hire with a competitive compensation package.
There expert team works around the clock to enhance the
high-quality education we provide to our pupils. Additionally,
there students will work on client locations, where they will
gain real experience in understanding and framing
customers' particular requirements.
b) Introduction of project

• The handwritten digit recognition is the ability of

computers to recognize human handwritten digits. It is a
hard task for the machine because handwritten digits
are not perfect and can be made with many different
flavors. The handwritten digit recognition is the solution
to this problem which uses the image of a digit and
recognizes the digit present in the image.
• The ability of computers to recognize human
handwritten digits is known as handwritten digit
recognition from sources such as paper documents,
images, touch-screen etc.
• In this project we will discover how to develop a deep
learning model to achieve high performance on the
handwritten digit recognition task using
the MNIST dataset and build a GUI App based
on Tkinter where, you can draw the digits (single as
well as multiple) and recognize it straight away by draw
a bounding box surrounding each digit.
• The digit recognition project aims to develop an advanced system that
can automatically identify handwritten digits (0-9) from image data
with high accuracy. This involves creating an algorithm capable of
handling the wide variability in handwriting styles, sizes, thicknesses,
and orientations, ensuring robust performance even in the presence of
noise or distortions.

• The project also focuses on scalability, enabling the model to process

large datasets efficiently and perform well in real-time applications.
User-friendliness is prioritized by designing an interface that allows
users to easily input digit images and receive accurate predictions,
making the technology accessible to non-experts. Furthermore, the
model's ability to generalize to new, unseen data is crucial, ensuring
reliable performance across diverse handwriting styles. The project also
considers integration potential, making the system adaptable to existing
applications such as automated form processing and educational tools.
By leveraging advanced techniques, including deep learning, this
project seeks to push the boundaries of digit recognition technology,
contributing to the automation and enhancement of tasks involving
handwritten digit processing.
c) Problem definition

• It is easy for the human to perform a task accurately by

practicing it repeatedly and memorizing it for the next
time. Human brain can process and analyse images
easily. Also, recognize the different elements present in
the images.
• The challenge in handwritten digit recognition is mainly
caused by the writing style variations of every single
individual. So, it is not easy for the machine to recognize
the handwritten digits accurately like the humans do.
Hence, robust feature extraction is very important to
improve the performance of machines.
1. Text Recognition: Develop an Optical Character
Recognition (OCR) system that can extract and digitize
text from images, including printed documents,
handwritten notes, and scanned images.
2. Object Recognition: Create a system that can
identify and classify specific objects or patterns in digital
images, such as logos, barcodes, or faces.
3. Accuracy: Ensure high accuracy in recognition,
minimizing errors related to text misinterpretation or
object misclassification.
4. Efficiency: Optimize the system to process images
quickly, enabling real-time or near-real-time recognition.
5. Versatility: Design the system to handle various
input formats, including different languages, font styles,
and image qualities.
6. User Interface: Provide an intuitive interface for
users to upload images and retrieve the recognized
output easily.
 Handling variations in handwriting and font styles.
 Dealing with noisy or low-quality images.
 Recognizing text or objects in complex backgrounds.

7.Evaluation Metrics:
 Accuracy: The percentage of correctly identified digits.
  Confusion Matrix: A table to visualize the performance of
the model in terms of true positives, false positives, true
negatives, and false negatives.
 Precision, Recall, and F1-Score: To measure the balance
between the precision (how many selected items are
relevant) and recall (how many relevant items are
selected).

d) Limitation of exiting company

1 Resource Constraints:
 The company might have constraints in terms of the latest
technological tools or software licenses, which could limit
the scope and capabilities of the project interns can work
on. Financial limitations could restrict the availability of
advanced hardware or software, affecting the
implementation of more sophisticated technologies.

2 Project Scope and Complexity:

Intern projects may be constrained to simpler tasks or less
complex systems due to company policies or client
requirements. This could limit the depth of experience
intern gain. Interns might be assigned predefined projects
with limited flexibility, reducing opportunities for creativity
or innovation.
3 Mentorship and Guidance:
Limited availability of experienced mentors could result in infrequent
feedback and guidance, affecting the learning curve for interns. Mentors
may be too busy with their own tasks, leaving insufficient time for thorough
mentoring of interns. There might be a lack of workshops or training
sessions that cover advanced topics or emerging technologies, hindering
skill development.
4 Limited Scope of Work:
Interns might be assigned repetitive, low-impact tasks such
as data entry, documentation, or basic testing. This restricts
exposure to complex and meaningful projects that could
provide valuable experience. Interns may only work on
specific aspects of a project, such as coding small modules
or handling predefined functions, without gaining a holistic
understanding of the project lifecycle.
5 Limited Interaction with Clients:
Interns may not have direct interactions with clients or
stakeholders, which can limit their understanding of client
requirements, feedback, and project management aspects.
Opportunities to participate in client meetings or
presentations might be restricted, affecting the interns'
ability to develop client-facing skills and understand client
dynamics.
6 Limited Time for Project Completion
 Interns might work on projects with short timelines, which
may not allow sufficient time for thorough analysis,
development, and refinement. Due to the short duration of
the internship, interns might not be able to see the
complete lifecycle of a project, affecting their understanding
of project management.

7 Insufficient Feedback:
Without regular feedback, interns may not understand how
to improve their performance or correct mistakes. This can
hinder their learning curve and development of necessary
skills. Interns might not be aware of whether they are
meeting the company’s expectations or if they need to
adjust their approach or work style. Insufficient feedback
can lead to a lack of motivation and engagement as interns
may feel their efforts are not being recognized or valued.
Without feedback, interns are more likely to repeat the same
mistakes, which can affect the quality of their work and the
efficiency of their projects.
8 Limited Access to Resources:
Inability to Complete Tasks: Limited access to necessary
tools, software, or hardware can hinder the intern’s ability to
complete projects and assignments effectively .resources
can lead to delays in project completion, as interns may
need to work around the limitations or wait for access to
essential tools. Interns may miss out on valuable hands-on
experience with industry-standard tools and technologies if
they do not have adequate access. Limited resources can
prevent interns from exploring new technologies or
methodologies, which can impede their learning and skill
development
e) objective of project
1. Text Recognition: Develop an Optical Character
Recognition (OCR) system that can extract and digitize
text from images, including printed documents,
handwritten notes, and scanned images.
2. Object Recognition: Create a system that can identify
and classify specific objects or patterns in digital images,
such as logos, barcodes, or faces.
3. Accuracy: Ensure high accuracy in recognition,
minimizing errors related to text misinterpretation or
object misclassification.
4. Efficiency: Optimize the system to process images
quickly, enabling real-time or near-real-time recognition.
5. Versatility: Design the system to handle various input
formats, including different languages, font styles, and
image qualities.
6. User Interface: Provide an intuitive interface for users
to upload images and retrieve the recognized output
easily.
7. Automated Recognition:
Developing an algorithm that can automatically identify
handwritten digits (0-9) from image data with high accuracy.
8. Improving Accuracy:
Achieving high classification accuracy across a wide range
of handwritten styles, including different sizes, thicknesses,
and orientations.
9. Scalability:
Ensuring the model can process large datasets and perform
well in real-time applications, making it scalable for practical
use.
10. Robustness:
Creating a system that can handle noise, variations in digit
style, and distortions in the input images, ensuring reliable
performance in diverse conditions.
11. User-Friendliness:
 Designing an interface or application where users
can easily input digit images (e.g., by scanning or
uploading) and receive accurate predictions, making
the technology accessible to non-experts.
12. Generalization:
1. Ensuring the model can generalize well to new,
unseen data, meaning it should perform well on
different handwriting styles that were not part of the
training data.
13. Integration Potential:
Making the model or system easy to integrate with
existing applications, such as automated form
processing, educational tools, or data entry systems.
14. Contribution to the Field:
Enhancing the current state of digit recognition by
experimenting with and implementing advanced
techniques, such as deep learning, to push the
boundaries of what can be achieved in this domain

2.System Analysis
a) Feasible Study
1.) System Overview
The system allows for booking and managing patient
appointments through a GUI. It includes functionalities for
adding, viewing, removing, and resetting appointment
records. The system uses SQLite for data storage and
Tkinter for the interface.
2.) Technical Study
 Programming Language: Python is suitable for this
type of application due to its ease of use and extensive
library support.
 Database: SQLite is a good choice for a small-scale
application like this due to its simplicity and ease of
integration. However, for larger-scale applications or
those requiring concurrent access, a more robust
database system might be needed.
 GUI Library: Tkinter is appropriate for basic GUI
applications and provides necessary widgets for user
interaction. For more complex UIs, other libraries like
PYQT or KIVY could be considered.

3.) Usability
Learnability: How easy is it for new users to start using the
digit recognition system effectively?
Assessment: Evaluate the time it takes for users to
understand and use the system for recognizing digits. This
can include both the process of inputting digits (e.g.,
handwriting, image upload) and interpreting the results.

Efficiency: How quickly can users perform tasks using the

digit recognition system?
Assessment: Measure the time required for users to
complete tasks such as digit input and recognition, and the
speed at which the system processes and returns results.
Efficiency can also be judged by how quickly users can
correct mistakes if the recognition is inaccurate.
Accuracy: How reliably does the system recognize digits and
provide correct outputs?
Assessment: Analyze the accuracy of the digit recognition
results across various types of inputs (e.g., different
handwriting styles, image quality). Track the system's error
rate and its impact on user trust and satisfaction.

4. Economic Feasibility:
 Cost Estimation: The project’s costs include
computational resources, potential cloud service fees,
and any necessary software licenses. If using cloud-
based platforms, expenses can vary based on usage, but
these are generally scalable and manageable within a
project budget.
 Cost-Benefit Analysis: The benefits of automating digit
recognition—such as reducing manual data entry,
increasing efficiency, and minimizing errors—can justify
the investment. For organizations processing large
volumes of handwritten data, the return on investment
(ROI) can be significant.
 Funding and Resources: The project may require initial
funding for hardware or cloud services, but many
educational or research grants are available for AI and
machine learning projects. Additionally, using open-
source tools can reduce software costs.
5. Market Feasibility:
 Demand and Applications: There is a strong market
demand for digit recognition systems across various
industries, including finance (check processing), postal
services (ZIP code recognition), and education (automated
grading). This demand ensures the project's relevance and
potential for adoption

Competitor Analysis: While several existing digit recognition

solutions are available, the project can differentiate itself by
focusing on specific use cases or improving upon existing
models in terms of accuracy, speed, or ease of integration.
6. Operational Feasibility:
 Development Workflow: The project can follow an
agile development methodology, allowing for iterative
progress, continuous testing, and user feedback
integration. This approach ensures flexibility and
adaptability as the project evolves.
 Collaboration Tools: To manage the project effectively,
collaboration tools such as GitHub for version control,
JIRA for task management, and Slack for communication
can be employed. These tools facilitate teamwork,
especially if the project involves multiple developers or
is distributed across different locations.
 Deployment Strategy: Once developed, the system
needs to be deployed in an environment where it can be
accessed by end-users. This could involve deploying the
model as a web service using platforms like Flask or
FastAPI, or integrating it into mobile applications. Cloud
platforms like AWS, Azure, or Google Cloud can provide
scalable deployment options with easy access to APIs.
 User Training and Support: Operational feasibility
also involves preparing end-users to interact with the
system. This could mean creating user guides,
conducting training sessions, or developing intuitive
interfaces that require minimal user training. Support
systems should also be in place to assist users in
troubleshooting any issues they encounter.
7. social Feasibility:
 Budgeting for Resources: It’s crucial to estimate the
costs of cloud computing services (e.g., AWS EC2
instances for training), storage for large datasets, and
any specialized software licenses that may be required.
Budget planning should also consider costs for ongoing
maintenance and potential scaling as the system is
adopted more widely.
  Risk Management: Identifying and mitigating financial
risks is essential. This includes risks like underestimating
the resources required for model training, unexpected
costs related to software licenses, or delays in project
timelines leading to budget overruns. Contingency
planning should be in place to address these risks.
 Economic Impact: The project’s economic feasibility
also depends on its potential to reduce costs in current
workflows. For example, automating manual data entry
tasks through digit recognition could significantly cut
down on labor costs and reduce the time spent on
processing handwritten documents. Calculating potential
savings versus project costs can help justify the
investment.

b) Hardware and Software requirements

Hardware Requirements
1. Processor:
o Minimum: Dual-core processor (2.0 GHz or higher)
o Recommended: Quad-core processor (3.0 GHz or
higher)
o Purpose: Handles multitasking and database
management tasks efficiently.
2. Memory (RAM):
o Minimum: 4 GB
o Recommended: 8 GB or higher
o Purpose: Ensures smooth operation and handles
larger data sets without lag.
3. Storage:
o Minimum: 256 GB SSD or HDD
o Recommended: 512 GB SSD or higher
o Purpose: For storing the SQLite database and other
project files.
4. Display:
o Minimum: 1366x768 resolution
o Recommended: 1920x1080 resolution (Full HD)
o Purpose: To comfortably display the GUI
components.
5. Other Peripherals:
o Keyboard, Mouse, and Monitor.
o Optional: Printer for report printing.

Software Requirements
1. Operating System:
o Minimum: Windows 7/8/10, macOS, or Linux
o Recommended: Windows 10/11 or the latest
version of macOS/Linux
 o Purpose: Platform to run the Python code and
manage databases.
2. Python:
o Version: Python 3.6 or higher
o Purpose: To write and execute the code for the
patient management system.
3. SQLite:
o Version: Included with Python (SQLite3)
o Purpose: Lightweight database to store patient
information and appointment data.
4. MySQL:
.MY SQL Server 8.0 or later.
. MySQL Workbench for database management
(optional).
5. TKINTER:
. Purpose: Standard Python library for creating the
GUI.
. Packages Needed: TK calendar for the date picker
functionality.
6. IDE/Code Editor:
. Recommended: PyCharm, VS Code, or any text
editor like Sublime Text
. Purpose: Writing and debugging the Python code.
7. Python Libraries:
o PYMYSQL for connecting Python to MySQL databases.
o TKINTER for building the GUI.
o TK calendar for the appointment system.
o datetime for handling date and time functions.
o sqlite3 if you are using SQLite for managing the appointment
system database.
o TTK for improved styling in TKINTER applications.

8. Database:
o MySQL Database for storing patient records.
 o SQLite (optional) for handling smaller or local databases.

9. Additional Tools:
o Optional: Git for version control, Virtual Environment

for managing dependencies.

o MySQL Workbench (optional) for managing MySQL databases.
o XAMPP or WAMP (optional) if you want to manage the MySQL
server locally.
3)Development Environment
a) Introduction of technology used
Python:
 Role: Python serves as the primary programming
language for the application. Its simplicity, readability,
and vast ecosystem make it an ideal choice for
developing a robust desktop-based application like this
one.
 Advantages:
o Extensive standard libraries and community support.
o Versatility, enabling use in various domains such as
web development, automation, data analysis, etc.
o Easy to integrate with other technologies and
databases.
2. SQLite (Database Management System):
 Role: SQLite is used as the lightweight relational
database to store patient information, user credentials,
and other data. It is embedded directly into the
application, making it an ideal choice for a desktop
application.
 Advantages:
o Zero configuration: No server setup required.
o Lightweight and fast, suitable for smaller projects.
o Fully ACID-compliant, ensuring reliable transactions.
3. TKINTER (GUI Library):
 Role: TKINTER is the standard GUI (Graphical User
Interface) library in Python, used to create the
application's front end. It provides widgets such as
buttons, labels, text boxes, and more to build a user-
friendly interface.
 Advantages:
o Built into Python, requires no additional installation.
o Simple and easy to learn, with a straightforward API.
 o Cross-platform compatibility, works on Windows,
macOS, and Linux.
4. PIL (Python Imaging Library)/Pillow:
 Role: PIL, specifically its modern fork, Pillow, is used to
handle image processing tasks such as opening,
manipulating, and displaying images in the GUI. In this
project, it is used to load and display background
images, icons, and buttons.
 Advantages:
o Supports a wide range of image formats (JPEG, PNG,
GIF, etc.).
o Easy to use for basic image manipulation tasks.
o Integrates well with TKINTER for image display in
applications.
5. Message box (from TKINTER):
 Role: The Message box module is used to display pop-up
messages in the application. It is used for showing
alerts, errors, and informational messages to the user,
such as invalid login credentials.
 Advantages:
o Provides built-in dialog boxes for user interaction.
o Simplifies error handling and user notifications.
o Easy to integrate with other TKINTER widgets.
6. Image Tk (from PIL/Pillow):
 Role: Image Tk is used to convert images into a format
that can be displayed in TKINTER widgets. It allows the
seamless integration of images in the TKINTER GUI,
enhancing the visual appeal of the application.
 Advantages:
o Simplifies the process of handling images in TKINTER.
o Allows the use of various image formats without the
need for additional conversion tools.
o Enhances the visual design of the application.
4.System Design
a) Modular description
Objective: Develop a system that can accurately recognize
handwritten digits from images using machine learning
techniques.
Application: This project can be applied in postal code
recognition, check processing, and digitizing handwritten
notes.

2. Data Collection and Preprocessing

Dataset: Use the MNIST dataset, which contains 60,000
training images and 10,000 testing images of handwritten
digits.
Data Preprocessing:
Normalization: Scale pixel values to the range [0, 1].
Reshaping: Adjust the shape of images to fit the input
format of the model (e.g., 28x28 pixels).
One-Hot Encoding: Convert the labels into a one-hot
encoded format.
3. Model Selection
Convolutional Neural Networks (CNN): Leverage CNNs,
which are well-suited for image data due to their ability to
capture spatial hierarchies.
Model Architecture:
Input Layer: Accepts the preprocessed image data.
Convolutional Layers: Extract features using filters.
Pooling Layers: Downsample the feature maps to reduce
dimensionality.
Fully Connected Layers: Combine the features to make a
final prediction.
Output Layer: Uses a softmax activation function to
output a probability distribution across the 10 digit classes.
4. Model Training
Training Setup:
Loss Function: Use cross-entropy loss for classification.
Optimizer: Employ Adam or SGD for optimization.
Batch Size: Set an appropriate batch size (e.g., 32 or 64).
Epochs: Determine the number of epochs to iterate over
the dataset (e.g., 10-20).
Validation: Use a portion of the training data as a
validation set to tune hyperparameters and avoid
overfitting.

5. Model Evaluation
Metrics:
Accuracy: Measure the percentage of correctly classified
digits.
Confusion Matrix: Visualize the performance across all
digit classes.
Precision, Recall, and F1-Score*: Evaluate the model's
performance in detail.
Testing: Test the model on the unseen test dataset to
evaluate its generalization capability.
Data flow Diagram

ER-DIAGRAM
5) Implementation and Testing
a) Implementation and Screenshot
 1. Import the libraries and load the dataset
First, we are going to import all the modules that we are going to need for training our model. The
Keras library already contains some datasets and MNIST is one of them. So we can easily import the
dataset and start working with it. The mnist.load_data() method returns us the training data, its
labels and also the testing data and its labels.

2. Pre-process the data

• The image data cannot be fed directly into the model so we need to perform some
operations and process the data to make it ready for our neural network. The dimension of
the training data is (60000,28,28). The CNN model will require one more dimension so we
reshape the matrix to shape (60000,28,28,1).

3.Create the model

• Now we will create our CNN model in Python data science project. A CNN model
generally consists of convolutional and pooling layers. It works better for
data that are represented as grid structures, this is the reason why CNN works
well for image classification problems. The dropout layer is used to deactivate
some of the neurons and while training, it reduces offer fitting of the model.
We will then compile the model with the Adadelta optimizer.
4. Train the model
• The model.fit() function of Keras will start the training of the model.
It takes the training data, validation data, epochs, and batch size.
• It takes some time to train the model. After training, we save the weights and
model definition in the ‘mnist.h5’ file.

5. Evaluate the model

• We have 10,000 images in our dataset which will be used to evaluate how good
our model works. The testing data was not involved in the training of the data
therefore, it is new data for our model. The MNIST dataset is well balanced so
we can get around 99% accuracy.
B) Testing
6.Conclusion

• In this project, we have successfully built a

Python deep learning project on handwritten
digit recognition app. We have built and
trained the Convolutional neural network which
is very effective for image classification
purposes. Later on, we build the GUI where we
draw a digit on the canvas then we classify the
digit and show the results.

7.Biblography
1.Greeksforgreeks
2.Research gate
3.Slide share
4.Google sited