Image filtering (gaussian_filter) in SciPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of Gaussian filter on a simple 2D array

What is the output of applying gaussian_filter with sigma=1 to the following 2D array?

SciPy

import numpy as np
from scipy.ndimage import gaussian_filter

arr = np.array([[0, 0, 0], [0, 10, 0], [0, 0, 0]])
result = gaussian_filter(arr, sigma=1)
print(result.round(2))

A

[[0.46 0.91 0.46]
 [0.91 1.81 0.91]
 [0.46 0.91 0.46]]

B

[[0.00 0.00 0.00]
 [0.00 10.00 0.00]
 [0.00 0.00 0.00]]

C

[[1.00 1.00 1.00]
 [1.00 10.00 1.00]
 [1.00 1.00 1.00]]

D

[[0.10 0.20 0.10]
 [0.20 2.00 0.20]
 [0.10 0.20 0.10]]

💡 Hint

Think about how Gaussian smoothing spreads the value around neighbors.

The Gaussian filter smooths the array by spreading the central high value to neighbors, reducing the peak and increasing nearby values. The output shows a blurred version with values less than 10 but spread around the center.

🧠 Conceptual

intermediate

Effect of increasing sigma in gaussian_filter

What happens to the output image when you increase the sigma parameter in gaussian_filter?

A The image colors invert as sigma increases.

B The image becomes sharper as sigma increases.

C The image size increases as sigma increases.

D The image becomes more blurred as sigma increases.

💡 Hint

Think about what smoothing means for an image.

Increasing sigma increases the spread of the Gaussian kernel, causing more smoothing and thus a blurrier image.

🔧 Debug

advanced

Identify the error in gaussian_filter usage

What error will this code raise when run?

SciPy

from scipy.ndimage import gaussian_filter

arr = [[1, 2], [3, 4]]
result = gaussian_filter(arr, sigma='two')
print(result)

A TypeError: sigma must be a number or sequence of numbers

B ValueError: sigma cannot be a string

C SyntaxError: invalid syntax

D No error, runs successfully

💡 Hint

Check the type expected for sigma parameter.

The sigma parameter must be a number or a sequence of numbers. Passing a string causes a TypeError.

❓ data_output

advanced

Number of unique values after gaussian_filter

Given this 1D array, how many unique values does the output have after applying gaussian_filter with sigma=0.5?

SciPy

import numpy as np
from scipy.ndimage import gaussian_filter

arr = np.array([0, 0, 10, 0, 0])
result = gaussian_filter(arr, sigma=0.5)
unique_count = len(np.unique(result))
print(unique_count)

A 1

B 2

C 3

D 5

💡 Hint

Consider the symmetry of the array: edge positions are equal, next-to-center equal, center different.

The output array has 3 distinct smoothed values because of symmetry: the two edge positions have the same value, the two adjacent to center have the same value, and the center is different.

🚀 Application

expert

Choosing sigma for edge preservation in image smoothing

You want to smooth an image using gaussian_filter but keep edges sharp. Which sigma value strategy is best?

A Use a very large sigma to blur edges heavily.

B Use a very small sigma (close to 0) to minimize smoothing.

C Use sigma=0 to disable filtering.

D Use sigma=5 to smooth only edges.

💡 Hint

Less smoothing means edges stay sharper.

Small sigma values apply minimal smoothing, preserving edges better than large sigma values which blur edges.