File size: 2,717 Bytes
561e703 2003d2f 561e703 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 |
---
title: Road Sign Recognition System
emoji: 🚦
colorFrom: red
colorTo: yellow
sdk: gradio
sdk_version: 4.44.0
app_file: app.py
pinned: false
license: mit
---
# 🚦 Road Sign Recognition System
## Overview
An intelligent computer vision application that identifies and classifies road signs from images using AutoGluon's MultiModal Predictor. This system helps with driver education, navigation assistance, and road safety analysis.
## Model Details
- **Model Type**: AutoGluon MultiModal Predictor
- **Architecture**: Deep learning with vision transformer backbone
- **Source Model**: [its-zion-18/sign-image-autogluon-predictor](https://huggingface.co/its-zion-18/sign-image-autogluon-predictor)
- **Task**: Multi-class image classification for road sign detection
- **Training**: AutoGluon automatic model selection and hyperparameter tuning
-
## Limitations
- **Not Enough Classes** Currently only have 2 classes (Stop and Speed limit)
## Features
### Input Methods
- **File Upload**: Support for JPG, PNG, and other common image formats
- **Webcam Capture**: Real-time photo capture for immediate classification
- **Clipboard Paste**: Quick image input from clipboard
- **Example Images**: Pre-loaded road sign samples for testing
### Output Information
- **Primary Classification**: Most likely sign type with confidence score
- **Top Predictions**: Alternative classifications with probabilities
- **Confidence Visualization**: Progress bars showing probability distribution
- **Detailed Analysis**: Interpretation of confidence levels
## How to Use
1. **Upload or Capture**: Provide an image of a road sign
2. **Automatic Processing**: Model analyzes the sign immediately
3. **View Results**: See the classification and confidence scores
4. **Try Examples**: Test with provided sample images
## Best Practices
- Use clear, well-focused images
- Ensure the entire sign is visible
- Avoid excessive angles or distortion
- Good lighting improves accuracy
- One sign per image works best
## Technical Specifications
- **Framework**: AutoGluon for automated machine learning
- **Backend**: PyTorch for deep learning operations
- **Deployment**: Hugging Face Spaces
- **Interface**: Gradio for user interaction
## Applications
- Driver education and training
- Navigation system enhancement
- Municipal sign inventory
- Road safety assessment
- Autonomous vehicle development
## Limitations
- Optimized for standard road signs
- Single sign detection per image
- May vary with weather conditions
- Region-specific signs may not be recognized
## Privacy
- No image storage after processing
- No personal data collection
- Server-side processing only
---
*Created for CMU 24-679 Course Assignment* |