YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
🛡️ VisionGuard AI: Logo & Watermark Detection Suite
VisionGuard is a production-ready system for the detection and removal of logos and watermarks from images, videos, and real-time feeds. Powered by YOLOv8 and built with a premium Streamlit UI.
🚀 How to Run the Project
Follow these steps to get the system running perfectly on your local machine.
1. Prerequisites
Ensure you have Python 3.9+ installed. This project does not require Docker.
2. Setup Environment
Open your terminal (PowerShell on Windows or Bash on Mac/Linux) and run:
# Navigate to the project directory
cd watermarklogo
# Create a virtual environment
python -m venv .venv
# Activate the virtual environment
# On Windows:
.\.venv\Scripts\activate
# On Mac/Linux:
source .venv/bin/activate
# Install dependencies
pip install -r requirements.txt
3. Generate Sample Data (Optional)
If you don't have a dataset yet, you can generate synthetic images to test the pipeline:
python scripts/generate_synthetic_dataset.py
4. Train the Model
To train the model on your dataset (or the synthetic one):
python src/train.py --data data.yaml --epochs 100 --imgsz 640 --batch 16
Note: A pre-trained model will be saved in models/best.pt.
5. Launch the Premium Dashboard
This is the main application interface.
streamlit run app/app.py
🛠️ Advanced Usage
Run a Detection Demo
To quickly verify that the detection logic is working perfectly:
python scripts/run_demo.py
Check the output in the outputs/ folder.
Evaluate Performance
To see metrics like mAP, Precision, and Recall:
python src/evaluate.py --weights models/best.pt --data data.yaml
📂 Project Structure
app/: Premium Streamlit UI.src/: Core detection, training, and removal logic.scripts/: Helper scripts for data generation and demos.models/: Stores your trained.ptweights.dataset/: Your images and labels.
© 2026 VisionGuard AI | Built with YOLOv8 & Streamlit