🧠 Image Classification with NN vs CNN (Fashion MNIST) - PyTorch

This project compares the performance of a fully connected Neural Network (NN) and a Convolutional Neural Network (CNN) on the Fashion MNIST dataset using PyTorch.

🎯 Objective

• Train two models (NN and CNN) on Fashion MNIST
• Evaluate and compare their accuracy
• Visualize sample predictions
• Analyze why CNN performs better for image classification

---

📦 Dataset

• Fashion MNIST (from torchvision.datasets)
  • 60,000 training images
  • 10,000 test images
  • 28x28 grayscale images
  • 10 clothing categories (T-shirt/top, Trouser, Pullover, Dress, etc.)

---

🧠 Models

1️⃣ Neural Network (NN)

• Fully connected (Dense) layers only
• Layers: Flatten → Dense(128) → Dense(64) → Output(10)

2️⃣ Convolutional Neural Network (CNN)

• Layers: Conv2D → ReLU → MaxPool → Conv2D → MaxPool → Flatten → Dense → Output(10)

---

🛠️ Setup Instructions

1. Clone the Project

git clone https://github.com/your-username/image-classification-nn-vs-cnn.git cd image-classification-nn-vs-cnn

2. Create Virtual Environment (optional but recommended)

python -m venv venv

Activate:

venv\Scripts\activate # Windows source venv/bin/activate # macOS/Linux

3. Install Requirements

pip install -r requirements.txt

4. Run the Project

python main.py

📈 Output Example Test accuracy for NN and CNN

5 sample predictions per model

Comparison of accuracy: Example:

objectivec Copy Edit NN Accuracy: 0.8555
CNN Accuracy: 0.9106
CNN performed better than NN by 5.51%

🔍 Why CNN Performs Better CNN captures spatial features using convolution filters.

CNN is translation invariant and parameter efficient for image data.

NN flattens image, losing spatial relationships.

✅ Requirements Python ≥ 3.7 (✅ works with 3.13)

PyTorch

TorchVision

Matplotlib