Finding Difference between Images using PIL

In image processing, finding the difference between two images is a crucial step in various applications. It is essential to understand the difference between the two images, which can help us in detecting changes, identifying objects, and other related applications. In this tutorial, we will explore how to find the difference between two images using the Python Imaging Library (PIL).

Installation

To use PIL, we need to install it using the pip package manager ?

pip install pillow

Basic Syntax

To find the difference between two images using PIL, we can use the ImageChops module. The ImageChops module provides various operations on images, including finding the difference between two images ?

from PIL import Image, ImageChops

# Open two images
img1 = Image.open('image1.jpg')
img2 = Image.open('image2.jpg')

# Find the difference
diff = ImageChops.difference(img1, img2)

# Display the difference
diff.show()

How It Works

The ImageChops.difference() function works by comparing each pixel in both images and calculating the absolute difference between their values. The result is a new image where:

• White pixels represent areas where the images differ significantly

• Black pixels represent areas where the images are identical

• Gray pixels represent minor differences between the images

Note: Both images should have the same dimensions for accurate comparison. If they differ in size, you should resize them first.

Example 1: Basic Image Difference

Let's create two simple images and find their difference ?

from PIL import Image, ImageDraw, ImageChops

# Create two similar images with slight differences
img1 = Image.new('RGB', (200, 200), 'white')
img2 = Image.new('RGB', (200, 200), 'white')

# Add shapes to both images
draw1 = ImageDraw.Draw(img1)
draw2 = ImageDraw.Draw(img2)

# Circle in first image
draw1.ellipse([50, 50, 150, 150], fill='red')

# Circle in second image (slightly different position)
draw2.ellipse([60, 60, 160, 160], fill='red')

# Find the difference
diff = ImageChops.difference(img1, img2)

# Convert to grayscale for better visibility
diff_gray = diff.convert('L')

print("Difference image created successfully")
print("Image mode:", diff_gray.mode)
print("Image size:", diff_gray.size)

Difference image created successfully
Image mode: L
Image size: (200, 200)

Example 2: Working with Real Images

Here's how to compare two actual image files with resizing and thresholding ?

from PIL import Image, ImageChops

# Create sample images for demonstration
img1 = Image.new('RGB', (300, 300), 'lightblue')
img2 = Image.new('RGB', (300, 300), 'lightblue')

# Add different elements to each image
from PIL import ImageDraw
draw1 = ImageDraw.Draw(img1)
draw2 = ImageDraw.Draw(img2)

draw1.rectangle([50, 50, 150, 150], fill='yellow')
draw2.rectangle([100, 100, 200, 200], fill='yellow')

# Resize images to ensure same dimensions
img1 = img1.resize((400, 400))
img2 = img2.resize((400, 400))

# Find the difference
diff = ImageChops.difference(img1, img2)

# Apply threshold to highlight significant differences
threshold = 50
diff_threshold = diff.point(lambda x: 0 if x < threshold else 255)

print("Original image size:", img1.size)
print("Difference image created with threshold:", threshold)

Original image size: (400, 400)
Difference image created with threshold: 50

Parameters and Methods

Common Use Cases

• Change Detection: Monitor surveillance footage to detect movement or changes in scenes

• Medical Imaging: Compare medical scans to identify tumors or track disease progression

• Quality Control: Inspect manufactured products for defects or variations

• Forensic Analysis: Compare evidence photos to reveal important details

• Image Alignment: Verify proper registration between multiple images

Key Points

• Images must have identical dimensions for accurate comparison

• Use resize() to match image sizes before comparison

• Apply thresholding with point() to highlight significant differences

• Convert to grayscale for better difference visualization

• The result image shows differences as bright pixels and similarities as dark pixels

Conclusion

PIL's ImageChops module provides an efficient way to find differences between images using the difference() method. This technique is valuable for change detection, quality control, and forensic analysis applications where identifying visual differences is crucial.