Object Detection with PyTorch (torchvision)

This is a minimal example that runs a pretrained Faster R-CNN (ResNet-50 FPN) on a local image and saves an annotated image with bounding boxes and labels.

Project structure

.
├── data/              # Place your input image(s) here
├── src/
│   ├── __init__.py
│   └── detect.py      # Main script
├── requirements.txt
├── README.md
└── venv/              # (optional) Virtual environment lives here

Setup

1. Create a virtual environment (recommended)

• Windows (PowerShell):

python -m venv venv
./venv/Scripts/Activate.ps1

• macOS / Linux:

python3 -m venv venv
source venv/bin/activate

2. Install dependencies

pip install -r requirements.txt

Usage

1. Put an image in the data/ folder, for example data/input.jpg.

2. Run the detector:

python -m src.detect --image data/input.jpg --output data/output.jpg --score-thresh 0.6

Adjust appearance of boxes and labels:

• Increase line width:

python -m src.detect --image data/input.jpg --output data/output.jpg --score-thresh 0.6 --line-width 10

3. Check data/output.jpg for the annotated result. If no detections are above the threshold, the original image will be saved and you will see a console message stating so.

Customization

• Change --score-thresh to control how many boxes are shown.
• Use a different output path (e.g., --output data/dog_boxes.jpg).

Preview test output