person_autocrop

This is a Python script to auto-detect and auto-crop a person in a image.

This script uses Tensorflow and SSD Mobilenet V2 (COCO) to recognize people in a photo, then crop the photo to the single person.

Specifications

• Windows and Linux command-line auto-install scripts
  • Creates automatically the Python virtual environment and install requirements
  • Downloads automatically the SSD Mobilenet V2 COCO model
• Supports GPU computing
• Images extension supported: .jpg, .jpeg, .png, .webp, .bmp
• Auto-crop a person if found
  • If there is more than one person in the photo, it will create more photos
  • Batch processing a input folder
• Option to save image in a specific format or save as input format

Requirements

• Python 3.10 installed (with Python-pip)

How to use

1. Download the repository
2. Just run the .bat or .sh scripts.
   • If Windows: run.bat
   • If Linux: run.sh (be sure to chmod +x the script)

Manual installation and use

If the automatic script does not work or if you want to use the python script directly.

1. First, create a folder (name of your choice) and open a terminal in it.
2. Create a Python virtual environment:

python -m venv venv

3. Activate the environment:

   • If Windows: ./venv/Scripts/activate
   • If Linux: source venv/bin/activate

4. Install the requirements:

pip install -r requirements.txt

5. Run the Python script person_autocrop.py with:

python person_autocrop.py