https://github.com/gsaini/opencv-use-cases
A Python-based project for removing image backgrounds using OpenCV, featuring both simple threshold-based and advanced GrabCut algorithms. Includes a user-friendly Streamlit web interface for interactive processing.
https://github.com/gsaini/opencv-use-cases
computer-vision image-processing opencv pillow pillow-library python straml
Last synced: 10 days ago
JSON representation
A Python-based project for removing image backgrounds using OpenCV, featuring both simple threshold-based and advanced GrabCut algorithms. Includes a user-friendly Streamlit web interface for interactive processing.
- Host: GitHub
- URL: https://github.com/gsaini/opencv-use-cases
- Owner: gsaini
- Created: 2025-10-23T01:39:40.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2025-10-29T03:14:44.000Z (8 months ago)
- Last Synced: 2026-06-14T23:34:31.982Z (10 days ago)
- Topics: computer-vision, image-processing, opencv, pillow, pillow-library, python, straml
- Language: Python
- Homepage:
- Size: 5.86 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Open CV Use Cases
## Remove Background from Images Using OpenCV
Quick, practical steps and two working approaches: (A) fast for simple/uniform backgrounds, (B) robust using GrabCut for general photos.
Prerequisites
- Python 3.x
- pip install opencv-python numpy
Method A — Simple background removal (uniform background)
1. Convert image to grayscale, blur, and threshold.
2. Find the largest contour (assumed foreground).
3. Create a mask, refine with morphology, apply mask and save as PNG with alpha.
Example:
```python
import cv2
import numpy as np
def remove_bg_simple(in_path, out_path):
img = cv2.imread(in_path)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
blur = cv2.GaussianBlur(gray, (5,5), 0)
_, th = cv2.threshold(blur, 0, 255, cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)
# keep largest contour
contours, _ = cv2.findContours(th, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
mask = np.zeros_like(th)
if contours:
c = max(contours, key=cv2.contourArea)
cv2.drawContours(mask, [c], -1, 255, -1)
# refine mask
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (7,7))
mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, kernel, iterations=2)
mask = cv2.GaussianBlur(mask, (5,5), 0)
b, g, r = cv2.split(img)
rgba = [b, g, r, mask]
out = cv2.merge(rgba, 4)
cv2.imwrite(out_path, out)
```
Method B — Robust removal using GrabCut
1. Initialize a mask (rectangle or rough mask) and run cv2.grabCut.
2. Convert the result to a binary mask, refine with morphology.
3. Apply the mask to the original and save with alpha.
Example:
```python
import cv2
import numpy as np
def remove_bg_grabcut(in_path, out_path, iter_count=5, rect_padding=10):
img = cv2.imread(in_path)
h, w = img.shape[:2]
mask = np.zeros((h, w), np.uint8)
# rectangle slightly inside the image
rect = (rect_padding, rect_padding, w-rect_padding*2, h-rect_padding*2)
bgdModel = np.zeros((1,65), np.float64)
fgdModel = np.zeros((1,65), np.float64)
cv2.grabCut(img, mask, rect, bgdModel, fgdModel, iter_count, cv2.GC_INIT_WITH_RECT)
# mask: 0/2 background, 1/3 foreground
mask2 = np.where((mask==2)|(mask==0), 0, 255).astype('uint8')
# refine
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5,5))
mask2 = cv2.morphologyEx(mask2, cv2.MORPH_CLOSE, kernel, iterations=2)
mask2 = cv2.GaussianBlur(mask2, (3,3), 0)
b, g, r = cv2.split(img)
rgba = cv2.merge([b, g, r, mask2])
cv2.imwrite(out_path, rgba)
```
Tips
- For complex scenes, combine user input (scribbles) with GrabCut by initializing mask values cv2.GC_FGD/cv2.GC_BGD.
- Adjust morphological kernel sizes and blur to control edge softness.
- Save output as PNG to preserve transparency.
- For batch processing, run these functions over your dataset and tune parameters per image type.
Streamlit app
This repository includes a simple Streamlit app `background_removal.py` that lets you upload an image, choose a background removal method (Simple or GrabCut), preview the result and download a PNG with transparency.
Run the app:
```bash
source .venv/bin/activate
pip install -r requirements.txt
streamlit run background_removal.py
```
Use the sliders when GrabCut is selected to tune iterations and rectangle padding.