Image-Classification-with-CNN

This project performs image classification on the CIFAR-10 dataset using Convolutional Neural Networks (CNNs). It explores different approaches including:

• Simple CNN
• Transfer Learning
• Data augmentation and Early Stopping
• Resnet

The main workflow is controlled via Jupyter notebooks, allowing you to train, evaluate, and visualize different models.

---

Project Structure

• README.md – This file.
• main.ipynb – Main notebook controlling the workflow. Execute models via specific notebook cells.
• data_organizer.py – Preprocesses and organizes CIFAR-10 data.
• plot_data.py – Visualizes sample images and class distributions.
• model_simple_cnn_test_1.py – Standard CNN model for CIFAR-10 - test case 1.
• model_simple_cnn_test_2.py – Standard CNN model for CIFAR-10 - test case 2.
• model_data_augmentation_test_1.py – CNN with data augmentation applied - test case 1.
• model_data_augmentation_test_2.py – CNN with data augmentation applied - test case 2.
• model_transfer_learning_test_1.py – CNN with Transfer Learning (one-hot encoding, early stopping) applied - test case 1.
• model_tf.py – Standard CNN model for CIFAR-10.
• model_resnet.py – CNN using a ResNet backbone.
• model_ahmad_f1.py – ???
• main_sofia.ipynb – Notebook focusing on training and evaluating the augmented model plus the Training Learning
• Presentation - Image Classification With CNN – Final Presentation
• Week 3 - Results - Parameters.xlsx – Keeping track of our experiments - not all are tracked here

---

Installation

1. Clone the repository:

git clone https://github.com/DigiZapi/Image-Classification-with-CNN.git
cd Image-Classification-with-CNN

Usage

You can handle everything directly from the main.ipynb notebook.
Currently, the notebook is set up to run each model in separate code blocks.

If you want to run an additional model:

1. Import your model at the top of the notebook or in a new cell:

from model_new_model import build_model  # Replace with your model file and function

2. Create a new code block in the notebook.

3. Run your model in that code block, e.g.:

model = build_model(input_shape=(32, 32, 3), num_classes=10)
model.fit(x_train, y_train, validation_data=(x_val, y_val), epochs=20, batch_size=64)