https://github.com/drissiomar98/image-classification-cifar10
A deep learning project comparing ANN and CNN architectures for classifying 32×32 color images across 10 categories in the CIFAR-10 dataset.
https://github.com/drissiomar98/image-classification-cifar10
cnn-classification deep-learning keras-tensorflow tensorflow
Last synced: 4 months ago
JSON representation
A deep learning project comparing ANN and CNN architectures for classifying 32×32 color images across 10 categories in the CIFAR-10 dataset.
- Host: GitHub
- URL: https://github.com/drissiomar98/image-classification-cifar10
- Owner: drissiOmar98
- Created: 2025-06-01T21:34:55.000Z (4 months ago)
- Default Branch: master
- Last Pushed: 2025-06-10T21:05:56.000Z (4 months ago)
- Last Synced: 2025-06-10T22:20:43.392Z (4 months ago)
- Topics: cnn-classification, deep-learning, keras-tensorflow, tensorflow
- Language: Jupyter Notebook
- Homepage:
- Size: 343 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# CIFAR-10 Image Classification with Deep Learning
[](https://opensource.org/licenses/MIT)
[](https://www.python.org/downloads/)
[](https://www.tensorflow.org/)
[](https://keras.io/)
[](https://opencv.org/)
[](https://scikit-learn.org/)
[](https://jupyter.org/)
[](https://numpy.org/)
[](https://pandas.pydata.org/)
[](https://matplotlib.org/)
[](https://colab.research.google.com/github/your-username/cifar10-classification/blob/main/CIFAR10_Classification.ipynb)
A deep learning project comparing ANN and CNN architectures for classifying 32×32 color images across 10 categories in the CIFAR-10 dataset.
## Table of Contents
- [Overview](#overview)
- [Key Features](#key-features)
- [Technologies Used](#technologies-used)
- [Results](#results)
- [Installation](#installation)
## Overview
This project implements and compares two approaches for image classification:
- **Artificial Neural Network (ANN)**: Baseline fully-connected model
- **Convolutional Neural Network (CNN)**: Advanced architecture featuring:
- Convolutional layers with ReLU activation
- Batch normalization
- Dropout regularization (30-50%)
- Data augmentation pipeline
- L2 weight regularization
The CIFAR-10 dataset contains 50,000 training and 10,000 test images across 10 object categories.
## Key Features
✔️ Comprehensive ANN vs CNN performance comparison
✔️ Advanced data augmentation pipeline
✔️ Real-time training visualization
✔️ Detailed classification metrics (precision/recall/F1)
✔️ Confusion matrix analysis
✔️ Regularization suite (Dropout, BatchNorm, L2)
✔️ Early stopping with model checkpointing
## Technologies Used
### Core Frameworks
| Technology | Purpose | Version |
|------------|---------|---------|
|  | Deep learning framework | 2.12+ |
|  | High-level API | Built-in |
|  | Model evaluation | 1.2+ |
### Data Processing
| Technology | Purpose |
|------------|---------|
|  | Array operations |
|  | Data analysis |
### Visualization
| Technology | Purpose |
|------------|---------|
|  | Training plots |
|  | Statistical visuals |
### Model Optimization
| Technology | Purpose |
|------------|---------|
|  | Real-time augmentation |
|  | Prevent overfitting |
## Results
### Performance Comparison
| Model | Train Accuracy | Test Accuracy | Parameters|
|-------|----------------|---------------|------------|
| ANN | 49.4% | 46.2% | 820,874 |
| CNN | **73.2%** | **72.4%** | 1,253,674 |
## Installation
```bash
# Clone repository
git clone https://github.com/drissiOmar98/Image-classification-CIFAR10.git
cd Image-classification-CIFAR10