A-Z Uppercase Handwritten Alphabet Recognition Classification Using CNN

This project focuses on building a Convolutional Neural Network (CNN) model using Tensorflow to detect uppercase letters in handwritten images. The goal is to accurately classify the images of handwritten uppercase letters from A to Z.

Members

This project is developed by team C23-PS363 of Bangkit Academy 2023 Capstone Project

Name	Bangkit ID	Learning Path	Github Profile
Ega Fernanda Putra	M038DSX0496	Machine Learning	Profile
Muhammad Raffi Priyadiantama	M038DSX0498	Machine Learning	Profile
Muhammad Naufal A.	A013DSX2909	Mobile Development	Profile
Danil Ardi	A013DSX0990	Mobile Development	Profile
Ridho Kartoni Pasaribu	C013DSX0978	Cloud Computing	Profile
Ruben Tricahya Boediono	C038DSX0600	Cloud Computing	Profile

Dataset

The dataset used for training the models is sourced from Kaggle and consists of handwritten alphabet images. The dataset contains 26 folders (A-Z) containing handwritten images in size 28x28 pixels, each alphabet in the image is centre fitted to 2020 pixel box and each image is stored as Gray-level

The Distribution of the Alphabet Letter is stated in below

Letter	Count	Letter	Count	Letter	Count
O	57825	S	48419	U	29008
C	23409	T	22495	P	19341
N	19010	A	13869	M	12336
L	11586	R	11566	E	11440
Y	10859	W	10784	D	10134
B	8668	J	8493	H	7218
X	6272	Z	6076	Q	5812
G	5762	K	5603	V	4182
F	1163	I	1120

this is the link to kaggle dataset A-Z Uppercase Handwritten Alphabet

Architecture Used in Model

Using Convolutional Neural Network (CNN) model to Classify the image, we can modify the model architecture based on requirements and in this project we use this architecture

Requirements Used in this Project

• Tensorflow
• Python 3
• Pandas
• Matplotlib
• Numpy

How To Contribute

• Clone this repository
• Install Dependencies based on the requirements above
• Prepare the Dataset as specified above link
• Customize the script, model architecture, and others as needed
• Run the script