Medium 0 solved 0% acceptance

Box Blur

Blur a grayscale image by replacing each pixel with the rounded average of its 3 by 3 neighborhood.

What you'll learn

  • Visit the 3 by 3 neighbors of a grid cell
  • Skip neighbors outside the grid
  • Write results to a new grid so originals are not overwritten

Problem

A box blur softens an image by replacing every pixel with the average of itself and its neighbors. Here the image is a grid of brightness values from 0 to 255. For each pixel, take the 3 by 3 block centered on it and average the pixels of that block that lie inside the image: 9 pixels in the middle, 6 along an edge, 4 in a corner (fewer for images only 1 row or column wide). Always average the original values, never values you have already blurred.

Round the average to the nearest whole number with halves rounded up. With integers only, for a block with sum s and count c that is (2 × s + c) / (2 × c) with integer division.

Implement box_blur(pixels).

Input

pixels: The image, one list per row. All rows have the same length.

Output

Return the blurred image as a grid of the same size. An empty image gives an empty grid.

Constraints

0 ≤ rows ≤ 50, 1 ≤ pixels per row ≤ 50 when there is at least one row

0 ≤ each pixel ≤ 255

Examples

Example 1
Input:
pixels = [
    [10, 20, 30],
    [40, 50, 60],
    [70, 80, 90]
]

Expected output:
expected = [
    [30, 35, 40],
    [45, 50, 55],
    [60, 65, 70]
]
Input:
std::vector<std::vector<long long>> pixels = {
    {10, 20, 30},
    {40, 50, 60},
    {70, 80, 90}
};

Expected output:
std::vector<std::vector<long long>> expected = {
    {30, 35, 40},
    {45, 50, 55},
    {60, 65, 70}
};
Input:
long[][] pixels = new long[][] {
    new long[] {10, 20, 30},
    new long[] {40, 50, 60},
    new long[] {70, 80, 90}
};

Expected output:
long[][] expected = new long[][] {
    new long[] {30, 35, 40},
    new long[] {45, 50, 55},
    new long[] {60, 65, 70}
};

Explanation

The corner 10 averages 10, 20, 40, 50 to 30. The center averages all nine values to 50. The top middle averages 10, 20, 30, 40, 50, 60 to 35.

Example 2
Input:
pixels = [
    [0, 255]
]

Expected output:
expected = [
    [128, 128]
]
Input:
std::vector<std::vector<long long>> pixels = {
    {0, 255}
};

Expected output:
std::vector<std::vector<long long>> expected = {
    {128, 128}
};
Input:
long[][] pixels = new long[][] {
    new long[] {0, 255}
};

Expected output:
long[][] expected = new long[][] {
    new long[] {128, 128}
};

Explanation

Each pixel's block holds both pixels, and (0 + 255) / 2 = 127.5 rounds up to 128, so [[128, 128]].

How to approach it

Try on your own first, then open one hint at a time.

Hint 1

Loop over row and column offsets from −1 to 1 around each pixel.

Hint 2

Count how many neighbors are inside the grid, because edges and corners have fewer.

Hint 3

Read from the original grid and write to a separate result grid.

Expected complexity

O(rows × columns) time, O(rows × columns) extra space

References

This challenge is adapted for CodeUtility. Follow the statement and constraints above.

The best passing solution of each coder. Faster, smaller and better-reviewed solutions rank higher.

No passing solutions yet. Solve it to take first place.

Learn from others

Community solutions

No published solutions yet. Solve this challenge to share the first one.