Image Classification with Transfer Learning

Project Description

This project demonstrates an advanced image classification model using transfer learning with a pre-trained ResNet50 model. The model is trained on the CIFAR-10 dataset, which consists of 60,000 32x32 color images in 10 different classes. This project covers various stages, including data preprocessing, model training, evaluation, and deployment using a Flask web application.

Features

1. Data Preprocessing and Augmentation: Includes advanced data augmentation techniques to enhance the training dataset.
2. Transfer Learning: Utilizes a pre-trained ResNet50 model for feature extraction and adds custom layers for classification.
3. Model Evaluation: Detailed evaluation using metrics such as accuracy, precision, recall, F1-score, and confusion matrix.
4. Model Explainability: Visualizes which parts of an image contribute most to the model's predictions using Grad-CAM.
5. Deployment: Deploys the trained model using a Flask web application for real-time image classification.

Installation

To get started, clone the repository and install the necessary dependencies.

git clone https://github.com/SlimenFellah/Image-Classification-Model-with-Transfer-Learning.git
cd image-classification-with-transfer-learning
pip install -r requirements.txt

Running the Project

Step 1: Exploratory Data Analysis (EDA)

Perform EDA to understand the dataset better and visualize the images and class distribution.

python eda.py

Step 2: Train the Model

Train the image classification model using the CIFAR-10 dataset. This step includes data augmentation, transfer learning, and hyperparameter tuning.

python model.py

Step 3: Evaluate the Model

Evaluate the trained model using various metrics and visualize the results.

python evaluation.py

Step 4: Deploy the Model

Deploy the model using a Flask web application. This allows you to upload images and get real-time classification results.

python deployment.py

The Flask app will be available at http://127.0.0.1:5000.

Usage

Using the Deployed Model

1. Start the Flask application:

   python deployment.py

2. Use a tool like curl or Postman to send a POST request to the /predict endpoint with an image file.

   curl -X POST -F "file=@path/to/your/image.jpg" http://127.0.0.1:5000/predict

   You should receive a JSON response with the predicted class of the image.

Project Structure

• main.py: Main script to run the entire project.
• eda.py: Script for exploratory data analysis.
• model.py: Script to define and train the model.
• evaluation.py: Script for model evaluation and visualization.
• deployment.py: Script to deploy the model using Flask.
• requirements.txt: File containing the list of dependencies.

Requirements

• torch
• torchvision
• Flask
• numpy
• matplotlib
• seaborn
• scikit-learn
• Pillow

Acknowledgements

This project utilizes the CIFAR-10 dataset, obtained from the torchvision library, and the ResNet50 model pre-trained on ImageNet, also from torchvision.

License

This project is licensed under the MIT License - see the LICENSE file for details.