🧠 PixelBrain – CNN Image Classification Model

📌 Project Overview

PixelBrain is a Convolutional Neural Network (CNN) model built using PyTorch to classify images from the CIFAR-10 dataset into 10 categories. It learns spatial patterns using convolutional layers and predicts classes efficiently.

---

🖼️ Model Architecture

---

🚀 Features

• Built with PyTorch
• CNN-based deep learning model
• Image preprocessing using torchvision.transforms
• Clean and modular architecture
• Supports training and testing

---

📊 Dataset

• CIFAR-10

• 60,000 images (32×32 RGB)

• 10 Classes:

  • Airplane ✈️
  • Car 🚗
  • Bird 🐦
  • Cat 🐱
  • Deer 🦌
  • Dog 🐶
  • Frog 🐸
  • Horse 🐎
  • Ship 🚢
  • Truck 🚚

---

🏗️ Model Architecture Details

🔹 Convolutional Layers

• Conv1: 3 → 32 filters (3×3) + ReLU + MaxPool
• Conv2: 32 → 64 filters (3×3) + ReLU + MaxPool
• Conv3: 64 → 128 filters (3×3) + ReLU + MaxPool

🔹 Fully Connected Layers

• Flatten: 128 × 4 × 4 → 2048
• FC1: 2048 → 256
• FC2: 256 → 10

---

🔄 Data Preprocessing

transform = transforms.Compose([
    transforms.ToTensor(),
    transforms.Normalize((0.5,0.5,0.5),(0.5,0.5,0.5))
])

---

🧪 Training Workflow

1. Load dataset using DataLoader
2. Forward pass
3. Compute loss
4. Backpropagation (loss.backward())
5. Update weights (optimizer.step())

---

⚙️ Installation

pip install torch torchvision matplotlib

---

▶️ Run the Project

python train.py

---

📈 Output

• Predicts 10 image classes
• Outputs probability distribution

---

🔮 Future Improvements

• Data augmentation
• Transfer learning (ResNet, VGG)
• Model optimization
• Deployment using Streamlit

---

👨‍💻 Author

Ishwar Sonawane Python Developer | Machine Learning Enthusiast

---

⭐ Project Name Meaning

PixelBrain "Understanding images at pixel level like a human brain"