| title | emoji | colorFrom | colorTo | sdk | app_port | tags | pinned | short_description | license | |
|---|---|---|---|---|---|---|---|---|---|---|
SourceCode |
π |
red |
red |
streamlit |
8501 |
|
false |
Paranox 2.0 Hackathon |
mit |
π
Date: November 16, 2025
π Event: TechXNinjas β PARANOX 2.0 (3-Month National Level Innovation + 24-Hour Build Hackathon)
π₯ Team: Team SourceCode
π Live Demo: https://paranox-sourcecode.hf.space
π§ CNN Model Training Notebook: https://colab.research.google.com/drive/1v-TD755AdnonZuIX0nXLbM2DvF281xiu?usp=sharing
SmartLane AI is an AI-powered Traffic Intelligence Platform designed to optimize 4-way traffic intersections using:
π YOLOv8 Object Detection
π Multimodal Ambulance/Emergency Vehicle Detection
π― CNN Deep Learning Classification
π Real-Time Traffic Analytics & Signal Time Optimization
It automatically prioritizes emergency vehicles (Ambulance) by extending green light duration and clearing the lane using a multi-method decision model:
| Detection Layers (Priority Order) |
|---|
| 1οΈβ£ Manual Override Control |
| 2οΈβ£ YOLO + Color Signature Analysis |
| 3οΈβ£ CNN Ambulance Classifier |
| 4οΈβ£ Red/White Pattern Recognition |
| 5οΈβ£ Emergency Text/OCR Pattern Detection |
π Live Web App: https://paranox-sourcecode.hf.space
π Upload four road images (North, East, South, West) β SmartLane AI detects vehicle density and prioritizes emergency lanes.
| Feature | Description |
|---|---|
| π Vehicle Detection | YOLOv8 counts Cars, Buses, Trucks, Motorcycles |
| π Emergency Recognition | CNN + YOLO + Color + Text Pattern Fusion |
| π¦ Adaptive Traffic Signal Control | AI-driven timing & clearance logic |
| π’ Emergency Priority Mode | Automatically extends green light (35s) |
| π Traffic Intelligence Dashboard | Vehicle matrix, charts, priority rankings |
| π₯ Export Reports | Download .txt traffic analysis summary |
| Category | Tools |
|---|---|
| AI / CV Models | YOLOv8, TensorFlow CNN |
| Framework | Streamlit |
| Language | Python |
| Visualization | Matplotlib, Pandas |
| Image Processing | OpenCV, PIL |
SMARTLANE-AI/
β
βββ app.py # Main Streamlit Application
βββ ambulance_cnn_final.keras # Trained CNN Model
βββ model/ # Additional model assets (optional)
β βββ classes.txt
βββ README.md
βββ requirements.txt
git clone https://github.com/lovnishverma/SMARTLANE-AI.git
cd SMARTLANE-AIpip install -r requirements.txtstreamlit run app.pyThe system includes a custom-trained CNN for Ambulance vs Non-Ambulance classification.
π Training Notebook (Google Colab): β‘ https://colab.research.google.com/drive/1v-TD755AdnonZuIX0nXLbM2DvF281xiu?usp=sharing
Model Summary:
- Input: 192Γ192 RGB Image
- Output: Softmax (Ambulance / Non-Ambulance)
- Accuracy: β 99.2% on Validation
| Condition | Green Time |
|---|---|
| Emergency Vehicle | β± 35 seconds |
| Normal Traffic Base | β± 5 seconds |
| Extra Time | β1 sec / 2 vehicles |
| Max Green | β³ 25 seconds |
PARANOX 2.0 is a 3-month national-level innovation event + 24-hour build hackathon where students BUILD Β· PITCH Β· WIN by converting their prototypes into real-world products. SmartLane AI was developed in this challenge under TechXNinjas.
| Member | Role |
|---|---|
| π Lovnish Verma | Lead Developer / ML Engineer |
| π§ Chandan Saroj | Computer Vision & CNN Training |
| β‘ Prateek Dhar Dwivedi | UI/UX & Streamlit Integration |
| π° Aman Choudhary | Deployment & Optimization |
Want to join or collaborate? PRs are welcome!
This project is licensed under MIT License. You are free to modify and use with attribution.
If you find this useful, please star the repo:
π https://github.com/lovnishverma/SMARTLANE-AI π
For research or collaboration inquiries:
π§ Email: [email protected] π GitHub: https://github.com/lovnishverma π LinkedIn: https://www.linkedin.com/in/lovnishverma/
Transforming traffic. Saving lives. Intelligent cities start here.