Image rotation and zoom in SciPy - Step-by-Step Execution

Concept Flow - Image rotation and zoom

Load Image Data

↓

Apply Rotation

↓

Apply Zoom

↓

Output Transformed Image

The image is first loaded, then rotated by a given angle, followed by zooming in or out, resulting in a transformed image.

Execution Sample

SciPy

from scipy.ndimage import rotate, zoom
import numpy as np

image = np.array([[1,2],[3,4]])
rotated = rotate(image, 90, reshape=False)
zoomed = zoom(rotated, 2)

This code rotates a 2x2 image by 90 degrees and then zooms it by a factor of 2.

Execution Table

Step	Action	Input Shape	Parameters	Output Shape	Output Snapshot
1	Load Image	-	-	(2, 2)	[[1, 2], [3, 4]]
2	Rotate Image	(2, 2)	angle=90, reshape=False	(2, 2)	[[2, 4], [1, 3]]
3	Zoom Image	(2, 2)	zoom=2	(4, 4)	[[2, 2, 4, 4], [2, 2, 4, 4], [1, 1, 3, 3], [1, 1, 3, 3]]

💡 All transformations applied; final image shape is (4,4)

Variable Tracker

Variable	Start	After Rotation	After Zoom
image	[[1, 2], [3, 4]]	[[1, 2], [3, 4]]	[[1, 2], [3, 4]]
rotated	N/A	[[2, 4], [1, 3]]	[[2, 4], [1, 3]]
zoomed	N/A	N/A	[[2, 2, 4, 4], [2, 2, 4, 4], [1, 1, 3, 3], [1, 1, 3, 3]]

Key Moments - 3 Insights

Why does the rotated image still have shape (2, 2) after a 90-degree rotation?

Because the input image is square (2x2), rotating by 90 degrees swaps rows and columns but the shape remains (2, 2), as shown in step 2 of the execution_table.

Why does the zoomed image shape become (4, 4) after zooming by 2?

Zooming by a factor of 2 doubles each dimension, so (2, 2) becomes (4, 4), as shown in step 3 of the execution_table.

What does the output snapshot represent after zooming?

It shows the pixel values repeated to enlarge the image, visually confirming the zoom effect, as seen in the last row of the execution_table.

Visual Quiz - 3 Questions

Look at the execution_table at step 2. What is the rotated image's shape?

A (1, 2)

B (4, 4)

C (2, 2)

D (3, 3)

💡 Hint

Check the 'Output Shape' column at step 2 in the execution_table.

At which step does the image shape change from (2, 2) to (4, 4)?

A Step 3

B Step 1

C Step 2

D No shape change

💡 Hint

Look at the 'Output Shape' column in the execution_table for each step.

If the zoom factor was changed to 3, what would be the new output shape after zooming?

A (9, 9)

B (6, 6)

C (3, 3)

D (2, 2)

💡 Hint

Zoom multiplies each dimension by the zoom factor; see variable_tracker for zoom effect.

Concept Snapshot

Image rotation and zoom with scipy.ndimage:
- Use rotate(image, angle, reshape=False) to rotate without changing shape.
- Use zoom(image, factor) to resize.
- Rotation keeps shape for square images when reshape=False.
- Zoom multiplies image dimensions.
- Output is a transformed numpy array.

Full Transcript

This visual execution trace shows how an image represented as a 2x2 numpy array is first rotated by 90 degrees using scipy.ndimage.rotate with reshape=False, which keeps the shape the same because the image is square. Then, the rotated image is zoomed by a factor of 2 using scipy.ndimage.zoom, which doubles each dimension, resulting in a 4x4 image. The execution table details each step's input and output shapes and shows the pixel values before and after transformations. The variable tracker follows the changes in variables image, rotated, and zoomed. Key moments clarify common confusions about shape changes during rotation and zoom. The quiz tests understanding of these transformations and their effects on image shape.

Practice

1. What does the function scipy.ndimage.rotate do to an image?

easy

A. It adds noise to the image.

B. It turns the image by a specified angle.

C. It crops the image to a smaller size.

D. It changes the image colors.

Solution

1. Step 1: Understand the function purpose

   scipy.ndimage.rotate is designed to rotate images by a given angle in degrees.

2. Step 2: Compare options with function behavior

   Only turning or rotating the image matches the function's purpose; other options describe different image operations.

3. Final Answer:

   It turns the image by a specified angle. -> Option B

4. Quick Check:

   Rotate = Turn image [OK]

Hint: Rotate means turn image by angle [OK]

Common Mistakes:

• Confusing rotate with crop or color change
• Thinking rotate changes image size
• Assuming rotate adds effects like noise

2. Which of the following is the correct way to rotate an image array img by 45 degrees using scipy?

easy

A. scipy.ndimage.rotate(img, 45)

B. scipy.ndimage.zoom(img, 45)

C. scipy.ndimage.rotate(45, img)

D. scipy.ndimage.zoom(45, img)

Solution

1. Step 1: Check function signatures

   scipy.ndimage.rotate takes the image first, then the angle as the second argument.

2. Step 2: Validate correct argument order

   scipy.ndimage.rotate(img, 45) correctly uses rotate(img, 45). Options C and D swap arguments incorrectly, and B uses zoom instead of rotate.

3. Final Answer:

   scipy.ndimage.rotate(img, 45) -> Option A

4. Quick Check:

   rotate(image, angle) correct syntax [OK]

Hint: Image first, angle second in rotate() [OK]

Common Mistakes:

• Swapping argument order
• Using zoom instead of rotate
• Passing angle before image

3. What will be the shape of the output image after applying scipy.ndimage.zoom(img, 2) if img.shape is (100, 100)?

medium

A. (50, 50)

B. (100, 100)

C. (200, 200)

D. (100, 200)

Solution

1. Step 1: Understand zoom factor effect

   A zoom factor of 2 doubles the size of each dimension of the image.

2. Step 2: Calculate new shape

   Original shape is (100, 100). Doubling each dimension gives (200, 200).

3. Final Answer:

   (200, 200) -> Option C

4. Quick Check:

   Zoom 2x doubles shape [OK]

Hint: Zoom factor multiplies each dimension [OK]

Common Mistakes:

• Confusing zoom with cropping
• Thinking zoom keeps shape same
• Mixing width and height dimensions

4. You run this code but get an error:

scipy.ndimage.rotate(45, img)

What is the problem?

medium

A. The image should be the first argument.

B. The angle should be the first argument.

C. The rotate function does not accept two arguments.

D. The angle must be in radians, not degrees.

Solution

1. Step 1: Check argument order for rotate()

   scipy.ndimage.rotate expects the image array as the first argument, then the angle.

2. Step 2: Identify error cause

   Passing angle first and image second causes a type error because the function tries to treat the number as an array.

3. Final Answer:

   The image should be the first argument. -> Option A

4. Quick Check:

   Image first, angle second in rotate() [OK]

Hint: Image must come before angle in rotate() [OK]

Common Mistakes:

• Swapping argument order
• Assuming angle must be radians
• Thinking rotate takes only one argument

5. You want to rotate an image by 90 degrees and then zoom it to half its size. Which code sequence correctly does this?

hard

A. rotated = scipy.ndimage.rotate(img, 0.5)
zoomed = scipy.ndimage.zoom(rotated, 90)

B. zoomed = scipy.ndimage.zoom(img, 0.5)
rotated = scipy.ndimage.rotate(zoomed, 90)

C. zoomed = scipy.ndimage.zoom(img, 90)
rotated = scipy.ndimage.rotate(zoomed, 0.5)

D. rotated = scipy.ndimage.rotate(img, 90)
zoomed = scipy.ndimage.zoom(rotated, 0.5)

Solution

1. Step 1: Understand operation order

   First rotate by 90 degrees, then zoom by 0.5 to reduce size by half.

2. Step 2: Check code correctness

   rotated = scipy.ndimage.rotate(img, 90)
   zoomed = scipy.ndimage.zoom(rotated, 0.5)

   correctly rotates img by 90, then zooms the rotated image by 0.5. Other options mix argument order or use wrong values.

3. Final Answer:

   rotated = scipy.ndimage.rotate(img, 90) zoomed = scipy.ndimage.zoom(rotated, 0.5) -> Option D

4. Quick Check:

   Rotate 90°, then zoom 0.5 [OK]

Hint: Rotate first, then zoom with correct factors [OK]

Common Mistakes:

• Swapping zoom and rotate order
• Using zoom factor as angle
• Passing wrong argument order