Tutorial: Flip and Rotate Images with Python Pillow

Flipping and rotating images are common operations in image manipulation, often used to correct orientation or create visually appealing effects.

With Python's Pillow library, these tasks are straightforward and efficient.

In this tutorial, we will cover:

1. Installing Pillow

Install the Pillow library using pip:

pip install pillow

2. Flipping Images

Flipping an image mirrors it along a specified axis, either horizontally or vertically.

2.1 Horizontal Flip

The transpose() method with Image.FLIP_LEFT_RIGHT flips the image horizontally.

from PIL import Image

# Open an image
image = Image.open("example.jpg")

# Flip horizontally
flipped_horizontal = image.transpose(Image.Transpose.FLIP_LEFT_RIGHT)

# Save and display the result
flipped_horizontal.save("flipped_horizontal.jpg")
flipped_horizontal.show()

2.2 Vertical Flip

The transpose() method with Image.FLIP_TOP_BOTTOM flips the image vertically.

# Flip vertically
flipped_vertical = image.transpose(Image.Transpose.FLIP_TOP_BOTTOM)

# Save and display the result
flipped_vertical.save("flipped_vertical.jpg")
flipped_vertical.show()

3. Rotating Images

Pillow allows rotation using both fixed angles and arbitrary angles.

3.1 Rotating at Fixed Angles

The transpose() method provides predefined options for 90°, 180°, and 270° rotations.

# Rotate 90 degrees
rotated_90 = image.transpose(Image.Transpose.ROTATE_90)
rotated_90.save("rotated_90.jpg")
rotated_90.show()

# Rotate 180 degrees
rotated_180 = image.transpose(Image.Transpose.ROTATE_180)
rotated_180.save("rotated_180.jpg")
rotated_180.show()

# Rotate 270 degrees (or -90 degrees)
rotated_270 = image.transpose(Image.Transpose.ROTATE_270)
rotated_270.save("rotated_270.jpg")
rotated_270.show()

3.2 Rotating by Arbitrary Angles

For arbitrary rotations, use the rotate() method. It rotates the image counterclockwise by a specified angle.

# Rotate by 45 degrees
rotated_45 = image.rotate(45, expand=True)

# Save and display the result
rotated_45.save("rotated_45.jpg")
rotated_45.show()

Options for rotate()

• expand=True: Ensures the entire rotated image is included in the output canvas.
• resample: Allows applying filters like Image.Resampling.BILINEAR for smoother results.

rotated_smooth = image.rotate(45, expand=True, resample=Image.Resampling.BILINEAR)
rotated_smooth.save("rotated_smooth.jpg")
rotated_smooth.show()

4. Combining Flip and Rotate

You can combine flipping and rotating for more complex transformations.

Example: Flip Horizontally and Rotate 90 Degrees

# Flip horizontally
flipped = image.transpose(Image.Transpose.FLIP_LEFT_RIGHT)

# Rotate 90 degrees
flipped_and_rotated = flipped.transpose(Image.Transpose.ROTATE_90)

# Save and display the result
flipped_and_rotated.save("flipped_and_rotated.jpg")
flipped_and_rotated.show()

5. Batch Processing

Process multiple images in a directory, flipping or rotating them.

Example: Batch Rotate and Save

import os
from PIL import Image

def batch_rotate(input_folder, output_folder, angle):
    os.makedirs(output_folder, exist_ok=True)

    for filename in os.listdir(input_folder):
        if filename.endswith((".jpg", ".png", ".jpeg")):
            img = Image.open(os.path.join(input_folder, filename))
            
            # Rotate the image
            rotated_img = img.rotate(angle, expand=True)
            
            # Save the rotated image
            rotated_img.save(os.path.join(output_folder, filename))
            print(f"Rotated and saved: {filename}")

# Rotate all images in 'input_images' folder by 45 degrees
batch_rotate("input_images", "output_images", 45)

6. Practical Examples

6.1 Flipping and Rotating for Image Augmentation

Image augmentation is useful in machine learning to increase dataset variety.

import random

def augment_image(image_path, output_folder):
    os.makedirs(output_folder, exist_ok=True)
    
    image = Image.open(image_path)
    
    # Randomly flip and rotate
    if random.choice([True, False]):
        image = image.transpose(Image.Transpose.FLIP_LEFT_RIGHT)
    if random.choice([True, False]):
        image = image.transpose(Image.Transpose.FLIP_TOP_BOTTOM)
    
    angle = random.randint(0, 360)
    augmented_image = image.rotate(angle, expand=True)
    
    # Save the augmented image
    augmented_image.save(os.path.join(output_folder, "augmented_image.jpg"))
    augmented_image.show()

# Example usage
augment_image("example.jpg", "augmented_images")

6.2 Rotating an Image with a Background Color

When rotating with expand=True, the empty space is transparent. Add a background color to fill the space.

def rotate_with_background(image_path, angle, background_color):
    image = Image.open(image_path)
    rotated_image = image.rotate(angle, expand=True, fillcolor=background_color)
    rotated_image.save("rotated_with_background.jpg")
    rotated_image.show()

# Example usage
rotate_with_background("example.jpg", 45, "white")

7. Summary

Key Methods

• Flipping:
  • Image.Transpose.FLIP_LEFT_RIGHT
  • Image.Transpose.FLIP_TOP_BOTTOM
• Rotating:
  • Image.Transpose.ROTATE_90, Image.Transpose.ROTATE_180, Image.Transpose.ROTATE_270
  • rotate(angle, expand=True)

Common Use Cases

• Correcting image orientation.
• Creating augmented datasets for machine learning.
• Generating creative effects for graphics or presentations.

Experiment with these examples to understand how flipping and rotating images can fit into your projects.