Deep Learning CNN Model for Street View House Numbers (SVHN) Dataset

Overview

This repository contains a deep learning CNN model implemented using TensorFlow and Keras to recognize house numbers from the Street View House Numbers (SVHN) dataset. The model achieves 95% accuracy on the test data.

Dataset

SVHN Dataset:

A real-world image dataset for developing machine learning and object recognition algorithms.

Dataset Size:

73257 training images, 26032 testing images, and 531131 additional images.

Image Size:

32x32 pixels.

Model Architecture

Convolutional Neural Network (CNN):

5 convolutional layers, 2 fully connected layers.

Activation Functions:

ReLU, Softmax.

Optimization Algorithm:

Adam Optimizer.

Loss Function:

Categorical Cross-Entropy.

Requirements

TensorFlow: 2.x Keras: 2.x NumPy: 1.x Matplotlib: 3.x Seaborn: 0.x Python: 3.x

Usage

Clone the repository:

git clone https://github.com/your-username/SVHN-CNN-Model.git Install requirements: pip install -r requirements.txt Download SVHN dataset: python download_svhn.py Train the model: python train_model.py Evaluate the model: python evaluate_model.py

Files

download_svhn.py: Script to download SVHN dataset. train_model.py: Script to train the CNN model. evaluate_model.py: Script to evaluate the trained model. model.py: CNN model architecture. utils.py: Utility functions for data preprocessing and visualization. requirements.txt: List of dependencies.

License

This project is licensed under the MIT License.

Contributing

Contributions are welcome! Please fork the repository, make changes, and submit a pull request.

Acknowledgments

SVHN dataset creators: TensorFlow and Keras developers.

Commit History

Initial commit: Implemented CNN model for SVHN dataset with data preprocessing and accuracy/loss visualization

Future Work

Implement data augmentation to increase diversity of training data. Experiment with different architectures and hyperparameters. Use transfer learning or pre-trained models. Implement more advanced techniques like attention mechanisms or batch normalization.