Image Classifier using CIFAR-10 Dataset

This project implements an image classification model using the CIFAR-10 dataset, a well-known dataset of 60,000 32x32 color images in 10 different classes. The model uses TensorFlow and includes a convolutional neural network (CNN) architecture. Additionally, it provides functionality for saving and loading the trained model and predicting new images.

Features 🚀

• Preprocessing of the CIFAR-10 dataset.

• CNN architecture for image classification.

• Visualization of training images with their labels.

• Saving and loading the trained model.

• Prediction on new images using the trained model.

Technologies Used

• Python

• TensorFlow/Keras

• NumPy

• OpenCV

• Matplotlib

Directory Structure

• ├── image_classifier_ml.py # Main script containing the implementation.

• ├── car.jpg # Sample image for prediction.

• ├── image_classifier.model # Saved trained model (generated after training).

Installation ⬇️

Clone the repository:

git clone https://github.com/yourusername/image-classifier-cifar10.git
cd image-classifier-cifar10

##Install dependencies: Ensure you have Python installed, then install the required libraries:

pip install numpy opencv-python matplotlib tensorflow

Run the script:

python image_classifier_ml.py

Usage/Examples

Dataset Overview:

• The CIFAR-10 dataset is automatically downloaded and preprocessed.

• The dataset includes the following 10 classes:

airplane, car, bird, cat, deer, dog, frog, horse, ship, truck.

Training model 🤖

• The model is trained on a subset of the CIFAR-10 dataset (20,000 images for training and 4,000 images for testing).

• The architecture includes convolutional, pooling, flattening, and dense layers.

• Visualizing the Data:

• The first 16 images from the training dataset are displayed with their respective labels for visualization.

Evaluating model 📋

After training, the model is evaluated on the testing dataset to calculate accuracy and loss.

Saving and Loading the Model

• The trained model is saved to image_classifier.model.

• The saved model can be loaded and used for future predictions without retraining.

Prediction

• The script includes functionality to predict the class of a new image (car.jpg).

• The image is read using OpenCV, processed, and passed through the model for prediction.

• Project Workflow

• Data Loading and Preprocessing:

CIFAR-10 dataset is normalized to scale pixel values between 0 and 1.

Model Architecture

• Convolutional layers for feature extraction.

• MaxPooling layers for down-sampling.

• Dense layers for classification with a softmax activation function.

Training:

• Optimizer: Adam.

• Loss Function: Sparse Categorical Crossentropy.

• Metrics: Accuracy.

• Evaluation: Calculates accuracy and loss on the test set.

• prediction: Accepts new images, preprocesses them, and predicts their class using the trained model.

Results

The model is trained for 10 epochs.

Loss and accuracy are displayed after evaluation.

Sample predictions are made on the provided car.jpg image.

Future Enhancements

• Data Augmentation: Improve model robustness by augmenting the dataset (e.g., rotations, flips).

• Advanced Architectures: Use state-of-the-art architectures like ResNet or EfficientNet.

• Real-time Predictions: Implement a live prediction pipeline using a webcam or live feed.

• Deployment: Deploy the model using Flask or FastAPI for serving predictions via a REST API.

Contributing

Contributions are welcome! If you find any issues or have suggestions for improvement, feel free to open an issue or submit a pull request.

License

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

Acknowledgments

• CIFAR-10 Dataset

• [TensorFlow Documentation] (https://www.tensorflow.org/)

• [OpenCV Documentation] (https://docs.opencv.org/)