VisionAid is a web-based scene detection application that uses ResNet-50 pre-trained on the Places365 dataset to identify and classify scenes in real-time using your webcam.
- Real-time scene detection using webcam
- Top 5 scene predictions with confidence scores
- Simple, user-friendly web interface
- Uses state-of-the-art ResNet-50 pre-trained model
- Backend: Flask (Python)
- Machine Learning: PyTorch, ResNet-50
- Frontend: HTML5, JavaScript
- Model: Places365 pre-trained weights
- Python 3.8+
- pip
- Webcam-enabled device
git clone https://github.com/Eng-M-Abdrabbou/Scene-Description-Python-ResNet-50.git
cd Scene-Description-Python-ResNet-50python -m venv venv
source venv/bin/activate pip install -r requirements.txtpython app.pyOpen your browser and navigate to http://localhost:5000
- The application uses a pre-trained ResNet-50 model from Places365
- When you click "Detect Scenes", it captures a frame from your webcam
- The image is processed and sent to the backend for scene classification
- Top 5 scene predictions are displayed with confidence scores
- Architecture: ResNet-50
- Dataset: Places365
- Total Scene Categories: 365
- Input Size: 224x224 pixels
- Preprocessing: Resize, Center Crop, Normalize
- Requires a webcam
- Accuracy depends on lighting and image quality
- Limited to 365 scene categories
Contributions are welcome! Please read the contributing guidelines before getting started.
This project is licensed under the MIT License - see the LICENSE file for details.
- Places365 Dataset
- ResNet-50 Model
- PyTorch Team
- Flask Framework
Mahmoud Abdrabbou - mahmoud.f.abdrabbou@gmail.com Project Link: https://github.com/Eng-M-Abdrabbou/Scene-Description-Python-ResNet-50.git