https://github.com/eshaasg/traffic-sign-classification

https://github.com/eshaasg/traffic-sign-classification

Last synced: 14 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/eshaasg/traffic-sign-classification
• Owner: eshaasg
• Created: 2024-12-11T09:11:43.000Z (23 days ago)
• Default Branch: main
• Last Pushed: 2024-12-11T15:24:42.000Z (23 days ago)
• Last Synced: 2024-12-11T15:32:11.205Z (23 days ago)
• Language: Jupyter Notebook
• Size: 669 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
![1sign](https://github.com/user-attachments/assets/0e8a7a78-7122-40d6-85f6-23159ddc4701)

# Traffic Signs Classification🚦

This project features a Convolutional Neural Network (CNN) designed to classify traffic signs into 43 distinct categories. It provides a reliable tool for understanding and categorizing traffic signs, which is essential for training datasets in applications such as autonomous driving and traffic analysis. By accurately identifying traffic signs, this classifier can aid in improving the efficiency of systems that rely on traffic data and serve as a foundational component for further developments in intelligent transportation solutions.

# Kaggle Dataset

🔗 [Traffic Signs images ](https://www.kaggle.com/code/yacharki/traffic-signs-image-classification-97-cnn)

The dataset used for this project is sourced from Kaggle. It consists of over 50,000 images, organized into 43 categories. Each category represents a specific traffic sign and contains between 210 to 2,500 images, providing a diverse and comprehensive dataset for training and evaluating the model.

# Approach

- Preprocesseed the data where the inages is normalized to have pixel values in the range[0,1]

- Built a CNN with 5 Convolutional layers, while pooling and flattening

- Added Dropout layers to prevent overfitting and enhance the model's ability to generalize across various traffic sign categories.

# Performance

- Accuracy : 98.8%

- Precision : 100.0%

- Recall : 99.0%

# Test Image Results

![23](https://github.com/user-attachments/assets/9335c23f-f85f-4d3d-8bf7-c0546e611a4e)

![2](https://github.com/user-attachments/assets/ca010de6-b8c7-4ab7-8cd3-cbb7f979221e)

# Links 

🔗 [Traffic Signs images ](https://www.kaggle.com/code/yacharki/traffic-signs-image-classification-97-cnn)