π¬ Smoke Detection with YOLOv11-Medium
This repository contains a custom-trained YOLOv11-Medium object detection model designed to detect cigarette smoke in images and videos. It is ideal for use in surveillance systems, public safety, and smoking zone enforcement.
π Model Performance
Metric | Value |
---|---|
Precision | 85.62% |
Recall | 76.92% |
[email protected] | 82.90% |
[email protected]:0.95 | 44.69% |
Evaluated on a custom Roboflow dataset using YOLOv11 medium variant trained for optimal balance of accuracy and speed.
π§ββοΈ Model Details
- Model: YOLOv11-Medium
- Task: Object Detection
- Classes:
cigarette
- Framework: Ultralytics YOLOv11
- Training Source: Roboflow Universe
π Dataset
- Name: Cigarette Smoke Detection
- Source: Roboflow Dataset
- Format: YOLO
- Annotations: Bounding Boxes
π‘ Usage
1. Load the Model
from ultralytics import YOLO
model = YOLO("path/to/best.pt") # Replace with your model path
2. Run Inference on an Image
results = model("your_image.jpg", save=True)
results.show()
3. Run Inference on a Video
results = model("your_video.mp4", stream=True)
for r in results:
r.show()
model.export(format="onnx")
π License
This model is available under the MIT License. Refer to the LICENSE file for more details.
π€ Author
Uploaded and maintained by: [Uppada Enos] Contact: [[email protected]]
π Model Hub
You can access and test this model directly via Hugging Face Spaces or API once published.
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
π
Ask for provider support
Model tree for Enos-123/smoking-detection
Base model
Ultralytics/YOLO11