Handwritten-Alphabets-Recognition

Tools and Techniques Used: This project utilizes Python programming language along with the following tools and techniques:

• OpenCV for image processing.
• Pandas for data manipulation.
• Matplotlib for data visualization.
• Scikit-learn for data preprocessing and model evaluation.
• Keras for building and training a Convolutional Neural Network (CNN).
• Google Colab as the development environment.
• Google Drive for data storage and retrieval.
• Kaggle for dataset acquisition and management.

Task of the Project:

1. Data Loading and Preprocessing:

   • Loaded a dataset of 370000 containing images of handwritten alphabets in CSV format.
   • Preprocessed the data, reshaped images, and prepared corresponding labels.

2. Model Development:

   • Designed a Convolutional Neural Network (CNN) using Keras for image classification.
   • Compiled the model with the Adam optimizer and categorical crossentropy loss.

3. Training and Evaluation:

   • Trained the CNN model on the prepared dataset, optimizing for accuracy.
   • Evaluated the model's performance on a test set, monitoring accuracy and loss.

4. Prediction and Visualization:

   • Applied the trained model to predict handwritten alphabets on an external image.
   • Visualized predictions along with the original image using OpenCV and Matplotlib.

Download the dataset from https://www.kaggle.com/datasets/sachinpatel21/az-handwritten-alphabets-in-csv-format