Python Basic Image Masking Techniques

Image masking is a powerful technique in computer vision. It allows you to isolate specific parts of an image. This guide covers basic masking methods in Python.

What Is Image Masking?

Image masking involves creating a binary image. The mask highlights regions of interest while hiding others. It's useful for object detection and image segmentation.

Masks work like stencils. White areas are visible, black areas are hidden. You can apply masks to extract or modify image sections.

Required Libraries

You'll need OpenCV and NumPy for image masking. Install them using pip if you haven't already.

# Install required libraries
pip install opencv-python numpy

Creating a Basic Mask

Let's start with a simple color-based mask. We'll detect red pixels in an image.

import cv2
import numpy as np

# Load image
image = cv2.imread('sample.jpg')

# Convert to HSV color space
hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)

# Define red color range
lower_red = np.array([0, 120, 70])
upper_red = np.array([10, 255, 255])

# Create mask
mask = cv2.inRange(hsv, lower_red, upper_red)

# Apply mask
result = cv2.bitwise_and(image, image, mask=mask)

# Save results
cv2.imwrite('mask.jpg', mask)
cv2.imwrite('result.jpg', result)

The cv2.inRange function creates the mask. cv2.bitwise_and applies it to the original image.

Masking with Thresholding

Thresholding converts grayscale images to binary masks. It's useful for simple segmentation tasks.

# Convert to grayscale
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

# Apply threshold
_, mask = cv2.threshold(gray, 127, 255, cv2.THRESH_BINARY)

# Invert mask if needed
mask = cv2.bitwise_not(mask)

Thresholding works well for high-contrast images. For more complex cases, try advanced segmentation techniques.

Combining Masks

You can combine multiple masks using logical operations. This creates more precise selections.

# Create two masks
mask1 = cv2.inRange(hsv, lower_red, upper_red)
mask2 = cv2.threshold(gray, 200, 255, cv2.THRESH_BINARY)[1]

# Combine masks
combined_mask = cv2.bitwise_or(mask1, mask2)

Use bitwise_and for intersection or bitwise_or for union of masks.

Applying Masks to Images

After creating a mask, apply it to modify your image. Here's how to replace masked areas.

# Create background (e.g., white)
background = np.full(image.shape, 255, dtype=np.uint8)

# Use mask to combine images
output = np.where(mask[:,:,None].astype(bool), image, background)

This technique is useful for creating collages or removing backgrounds.

Edge Detection Masks

Edge detection can help create masks for object boundaries. The Canny algorithm works well.

# Detect edges
edges = cv2.Canny(gray, 100, 200)

# Dilate edges to make thicker mask
kernel = np.ones((3,3), np.uint8)
edge_mask = cv2.dilate(edges, kernel, iterations=1)

Edge masks are great for highlighting contours. Combine them with other masks for better results.

Practical Applications

Image masking has many real-world uses:

• Object removal or isolation
• Background replacement
• Selective color adjustment
• Creating image mosaics

Conclusion

Python makes image masking accessible through OpenCV and NumPy. Start with simple color or threshold masks. Then combine techniques for complex results.

Remember to experiment with different approaches. Masking is often the first step in image processing pipelines.

For more advanced techniques, explore our image analysis guide.