Easy and fast way to remove background from video on Windows 11

BGSLibrary is designed for background subtraction . It analyzes a video to identify and separate a moving foreground (like a person or car) from a static background scene . This foreground can then be isolated. To learn how to remove background from video using BGSLibrary, you typically follow these steps: install the pybgs Python wrapper, run a simple script, and then apply the chosen algorithm to isolate the desired foreground elements.

The most straightforward way to use BGSLibrary on Windows is through its Python wrapper, pybgs. You will need Python, a few packages, and a C++ compiler like Visual Studio installed.

1. Install Prerequisites.

2. Install pybgs.

Once your prerequisites are ready, open the Command Prompt or PowerShell in your project folder, activate your virtual environment, and run the following command:

bash

pip install pybgs

This command installs the core BGSLibrary along with its Python bindings .

A Practical Python Example

After installation, you can create a Python script to process a video file. This script uses the SuBSENSE algorithm, which is a robust choice for many scenarios .

python

import cv2

import pybgs as bgs

 

# Open the video file

captured_video = cv2.VideoCapture("path/to/your/video.mp4")

if not captured_video.isOpened():

print("Error: Could not open the video.")

exit(0)

 

# Instantiate the background subtraction algorithm

background_subtr_method = bgs.SuBSENSE()

 

while True:

retval, frame = captured_video.read()

if not retval:

break

 

# Resize for faster processing (optional)

frame = cv2.resize(frame, (640, 360))

 

# Pass the frame to the subtractor to get the foreground mask

foreground_mask = background_subtr_method.apply(frame)

# Obtain the background model

img_bgmodel = background_subtr_method.getBackgroundModel()

 

# Display the results

cv2.imshow("Original Frame", frame)

cv2.imshow("Foreground Mask (Background Removed)", foreground_mask)

cv2.imshow("Background Model", img_bgmodel)

 

if cv2.waitKey(10) == 27: # Press 'ESC' to exit

break

 

captured_video.release()

cv2.destroyAllWindows()

In this example, foreground_mask is a black-and-white image where white areas represent the detected foreground (the part of the video that is moving), which is the first step in isolating it. While this doesn't automatically produce a video with a transparent or replaced background, it gives you the necessary mask to start. This is generally how to remove background from video using the pybgs library.