Batch Image Segmentation with Python and PaddlePaddle

Ever wanted to extract a person from a photo and paste them elsewhere, even if you are far away? While Photoshop’s Magic Wand works for professionals and mobile apps help casual users, they handle only one image at a time and can be slow for complex pictures.

This article introduces a third approach: using Python to perform one‑click batch background removal.

Preparation

The backbone of this solution is PaddlePaddle (also known as “飞桨”), an open‑source deep‑learning platform that abstracts the low‑level framework so you can focus on creative tasks. Install the CPU version via pip:

python -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple

After installation, verify it in a Python session:

import paddle.fluid paddle.fluid.install_check.run_check()

If the output contains “Your Paddle works well on SINGLE GPU or CPU”, the installation succeeded.

Next, install paddlehub to access pre‑trained models:

pip install -i https://mirror.baidu.com/pypi/simple paddlehub

Code Implementation

The workflow is simple: import the module, load the model, collect image files, and call the segmentation API.

import os, paddlehub as hub huseg = hub.Module(name='deeplabv3p_xception65_humanseg') # load model path = './imgs/' # image directory files = [path + i for i in os.listdir(path)] # list of image files results = huseg.segmentation(data={'image': files}) # perform segmentation

Place your source images in an imgs folder next to the script. After running, the segmented images are saved in a humanseg_output directory with the same filenames but in PNG format.

Running the program on five sample pictures produces five output files where the detected persons are isolated on a white background, demonstrating that the method works for both single and multiple subjects.

Although some fine details may still have minor artifacts, the results are generally satisfactory and provide a handy tool for anyone who frequently needs to cut out people from photos.

In summary, using PaddlePaddle and a few lines of Python code enables fast, batch‑style background removal, freeing users from manual editing and adding a useful trick to any developer’s toolbox.