🖼️ CIFAR-10 Image Classification using CNN

📌 Project Overview

This project demonstrates image classification using a Convolutional Neural Network (CNN) built with TensorFlow/Keras.
The model is trained on the CIFAR-10 dataset, which contains 60,000 color images (32×32 pixels) in 10 different classes.

It is a beginner-friendly deep learning project that helps understand how CNNs work for computer vision tasks.

---

🎯 Objectives

• Learn how to preprocess and normalize image datasets.
• Build a custom CNN model from scratch.
• Train the model using supervised learning.
• Evaluate accuracy on unseen test data.
• Provide a foundation for more advanced image classification tasks.

---

🗂️ Dataset – CIFAR-10

• Training set: 50,000 images
• Test set: 10,000 images
• Image size: 32×32 pixels (RGB)
• Classes (10):
  1. Airplane ✈️
  2. Automobile 🚗
  3. Bird 🐦
  4. Cat 🐱
  5. Deer 🦌
  6. Dog 🐶
  7. Frog 🐸
  8. Horse 🐎
  9. Ship 🚢
  10. Truck 🚚

---

🧠 Model Architecture

The CNN model is built using the Keras Functional API:

1. Input Layer: Shape (32, 32, 3)

2. Conv Block 1:

   • Conv2D (32 filters, 3×3, ReLU) → BatchNorm
   • Conv2D (32 filters, 3×3, ReLU) → BatchNorm
   • MaxPooling2D (2×2)
   • Dropout (0.2)

3. Conv Block 2:

   • Conv2D (64 filters, 3×3, ReLU) → BatchNorm
   • Conv2D (64 filters, 3×3, ReLU) → BatchNorm
   • MaxPooling2D (2×2)
   • Dropout (0.2)

4. Fully Connected Layers:

   • Flatten
   • Dense (1024 neurons, ReLU) → Dropout (0.2)
   • Dense (10 neurons, Softmax) → Output

---

⚙️ Techniques Used

• Supervised Learning
• Convolutional Neural Networks (CNNs)
• Batch Normalization (stabilizes learning)
• Dropout (reduces overfitting)
• MaxPooling (downsamples feature maps)
• Softmax Activation (multi-class classification)
• Adam Optimizer (adaptive learning rate)
• Sparse Categorical Crossentropy Loss

---

📊 Training

• Epochs: 50
• Optimizer: Adam
• Loss Function: Sparse Categorical Crossentropy
• Metric: Accuracy

---

📈 Results

• The model achieves around 70–80% test accuracy (depending on training).
• Accuracy can be improved with:
  • More training epochs
  • Data augmentation
  • Pre-trained models (VGG, ResNet, etc.)

---

🌐 Live Web App

👉 Click here to try the app

---

🚀 How to Run

1. Clone this repository:

   git clone https://github.com/your-username/cifar10-cnn.git
   cd cifar10-cnn

2. Install dependencies:

   pip install tensorflow numpy matplotlib

3. Run the notebook:

   jupyter notebook Biginner_level.ipynb

---

📌 Future Improvements

• Use Data Augmentation to make the model more robust.
• Try transfer learning with pretrained models like ResNet, VGG16, MobileNet.
• Deploy the model using Flask/Streamlit for a web demo.

---

🏆 Skills Gained

• TensorFlow/Keras
• CNNs for Computer Vision
• Image Preprocessing
• Model Training & Evaluation
• Overfitting Prevention (Dropout, BatchNorm)

---

📬 Author

👤 Syed Nabeel Ahmed

• GitHub
• LinkedIn

---

✨ This project is a beginner-friendly introduction to Deep Learning for Computer Vision using CNNs.