PIL Image Filters

Conceptual Anchor

A Filter Is a Transformation

Think about a photo-editing app. A filter takes an input image, applies a rule, and produces a new output image. Some rules look at neighbouring pixels (blur, sharpen, edges); others modify each pixel directly (negative, grayscale, sepia, warm tone).

Image.open() Load the source image

filter / enhance Apply a transformation

compare Inspect the result

save() Write a new image file

Built-in Filters — Visual Reference

The easiest way to understand an image filter is to compare the same source image after different transformations. All examples below use the same generated scene.

Original

No filter applied.

img

Gaussian Blur

Smooths local detail.

ImageFilter.GaussianBlur(4)

Sharpen

Emphasises local edges.

ImageFilter.SHARPEN

Find Edges

Highlights strong intensity changes.

ImageFilter.FIND_EDGES

Emboss

Produces a raised relief-like effect.

ImageFilter.EMBOSS

Smooth

Reduces small variations more gently.

ImageFilter.SMOOTH

Filter	Typical purpose	Python
`BLUR`	General blur / smoothing	`img.filter(ImageFilter.BLUR)`
`GaussianBlur(radius)`	Adjustable blur strength	`img.filter(ImageFilter.GaussianBlur(5))`
`SHARPEN`	Make edges more pronounced	`img.filter(ImageFilter.SHARPEN)`
`CONTOUR`	Extract contour-like lines	`img.filter(ImageFilter.CONTOUR)`
`DETAIL`	Enhance fine detail	`img.filter(ImageFilter.DETAIL)`
`EDGE_ENHANCE`	Strengthen edges	`img.filter(ImageFilter.EDGE_ENHANCE)`
`FIND_EDGES`	Detect edge-like intensity changes	`img.filter(ImageFilter.FIND_EDGES)`
`EMBOSS`	Create a relief / emboss effect	`img.filter(ImageFilter.EMBOSS)`
`SMOOTH`	Light smoothing	`img.filter(ImageFilter.SMOOTH)`

Interactive Filter Lab

Choose a filter to see the same image transformed immediately. The browser preview approximates the visual idea; the code shown below is the Pillow command you would use in Python.

Effect Original image Python img = Image.open("photo.jpg")

Using ImageFilter

1 Apply and Save a Built-in Filter

from PIL import Image, ImageFilter

img = Image.open("photo.jpg")

blurred = img.filter(
    ImageFilter.GaussianBlur(4)
)

blurred.save("photo_blurred.png")

2 Chain Filters

from PIL import Image, ImageFilter

img = Image.open("photo.jpg")

result = img.filter(ImageFilter.SHARPEN)
result = result.filter(ImageFilter.FIND_EDGES)

result.save("sketch_effect.png")

Order Matters

A filter changes the input for the next filter. SHARPEN → FIND_EDGES does not necessarily produce the same result as FIND_EDGES → SHARPEN.

Image Enhancement

ImageEnhance adjusts the appearance of the whole image. A factor of 1.0 keeps the original. Lower and higher factors decrease or increase the selected property.

Try the factors

Brightness 1.00

Contrast 1.00

Color 1.00

Sharpness 1.00

brightness = ImageEnhance.Brightness(img).enhance(1.00) contrast = ImageEnhance.Contrast(brightness).enhance(1.00) color = ImageEnhance.Color(contrast).enhance(1.00) result = ImageEnhance.Sharpness(color).enhance(1.00)

Enhancer	0.0	1.0	> 1.0
`Brightness`	Black	Original	Brighter
`Contrast`	Flat gray-like result	Original	More contrast
`Color`	Grayscale	Original	More colour saturation
`Sharpness`	Blurred	Original	Sharper

3 Increase Brightness

from PIL import Image, ImageEnhance

img = Image.open("photo.jpg")

enhancer = ImageEnhance.Brightness(img)
bright = enhancer.enhance(1.5)

bright.save("bright.png")

Why Do Blur and Edge Filters Look at Neighbours?

Many built-in filters do not change a pixel independently. They examine a small neighbourhood around the pixel and combine those values. This small matrix is commonly called a kernel.

3 × 3 Neighbourhood

P

X

P

The new value of centre pixel X can depend on the surrounding pixels P.

Conceptual difference

Blur	makes nearby values more similar
Sharpen	emphasises differences
Edge detection	looks for strong local changes

Custom Filters — Pixel Mathematics

A custom filter can read a pixel (r, g, b), calculate new channel values and write the result back. Use the controls below to see exactly what happens to one pixel.

Input Pixel

R 142

G 87

B 201

Input (142, 87, 201)

Output (113, 168, 54)

Formula

Worked Custom Filters

4 Manual Grayscale

from PIL import Image

img = Image.open("photo.jpg").convert("RGB")
w, h = img.size

for x in range(w):
    for y in range(h):
        r, g, b = img.getpixel((x, y))

        gray = int(
            0.299 * r +
            0.587 * g +
            0.114 * b
        )

        img.putpixel(
            (x, y),
            (gray, gray, gray)
        )

img.save("manual_gray.png")

5 Negative

from PIL import Image

img = Image.open("photo.jpg").convert("RGB")
w, h = img.size

for x in range(w):
    for y in range(h):
        r, g, b = img.getpixel((x, y))

        img.putpixel(
            (x, y),
            (255 - r, 255 - g, 255 - b)
        )

img.save("negative.png")

6 Sepia

from PIL import Image

img = Image.open("photo.jpg").convert("RGB")
w, h = img.size

for x in range(w):
    for y in range(h):
        r, g, b = img.getpixel((x, y))

        tr = min(255, int(0.393*r + 0.769*g + 0.189*b))
        tg = min(255, int(0.349*r + 0.686*g + 0.168*b))
        tb = min(255, int(0.272*r + 0.534*g + 0.131*b))

        img.putpixel((x, y), (tr, tg, tb))

img.save("sepia.png")

7 Threshold

Convert the pixel to an intensity. If that intensity is above the threshold, make it white; otherwise make it black.

65 < 128 → BLACK

threshold = 128

205 > 128 → WHITE

from PIL import Image

img = Image.open("photo.jpg").convert("RGB")
w, h = img.size

threshold = 128

for x in range(w):
    for y in range(h):
        r, g, b = img.getpixel((x, y))

        average = (r + g + b) // 3

        if average > threshold:
            new_color = (255, 255, 255)
        else:
            new_color = (0, 0, 0)

        img.putpixel((x, y), new_color)

img.save("threshold.png")

8 Warm Filter

from PIL import Image

img = Image.open("photo.jpg").convert("RGB")
w, h = img.size

for x in range(w):
    for y in range(h):
        r, g, b = img.getpixel((x, y))

        new_r = min(255, r + 20)
        new_g = g
        new_b = max(0, b - 20)

        img.putpixel(
            (x, y),
            (new_r, new_g, new_b)
        )

img.save("warm.png")

9 Keep Only the Red Channel

from PIL import Image

img = Image.open("photo.jpg").convert("RGB")
w, h = img.size

for x in range(w):
    for y in range(h):
        r, g, b = img.getpixel((x, y))

        img.putpixel(
            (x, y),
            (r, 0, 0)
        )

img.save("red_only.png")

Keep RGB Values in Range

Standard 8-bit RGB channels use values from 0 to 255. Custom formulas may produce values outside this range, so clamp values before writing them back.

10 Clamp a Calculated Value

# Keep a number between 0 and 255
new_r = min(255, max(0, calculated_r))
new_g = min(255, max(0, calculated_g))
new_b = min(255, max(0, calculated_b))

A Small Helper Function

For larger projects you can write def clamp(v): return min(255, max(0, v)) and reuse it for every channel.

Further Viewing

Blurring, Sharpening and Smoothing with Pillow

A short visual walkthrough from Real Python showing built-in ImageFilter operations and comparisons.

Watch

Common Pitfalls

1. filter() Returns a New Image

img.filter(...) does not replace img. Store the result: blurred = img.filter(ImageFilter.BLUR).

2. Unexpected Image Mode

If your custom filter expects r, g, b, first use img = img.convert("RGB").

3. Values Outside 0–255

A custom formula can produce negative values or values above 255. Clamp channel values before calling putpixel().

4. Slow Pixel-by-Pixel Code

Nested for loops are excellent for learning how filters work, but large images may process slowly. Resize a test image while developing.

5. Overwriting Your Original File

During practice, save results under a new filename such as photo_blur.png instead of replacing the source file.

Practice Tasks

Remember

Which Pillow module contains BLUR, SHARPEN and FIND_EDGES?

Understand

Explain the difference between a neighbour-based filter such as blur and a custom negative filter that changes each pixel independently.

Apply

Load an image, apply GaussianBlur(5), and save the result as blurred.png.

Apply

Use ImageEnhance.Brightness to create one version at factor 0.6 and another at 1.6.

Analyze

A pixel is (220, 80, 30). Calculate the result of the negative filter (255-r, 255-g, 255-b).

Analyze

A warm filter adds 50 to red. Explain why min(255, r + 50) is safer than simply using r + 50.

Create

Design your own custom filter using a clear mathematical rule. Explain what happens to R, G and B, then apply it to every pixel.

Create

Build a mini filter pipeline that produces three files from one source: a blurred image, an edge image and your own custom colour effect.

Self-Check Quiz

Q1: Which module provides BLUR, SHARPEN and FIND_EDGES?

ImageFilter from Pillow.

Q2: What does enhancement factor 1.0 mean?

It returns the original appearance for that enhancement property.

Q3: What does ImageEnhance.Color(...).enhance(0) produce?

A grayscale / black-and-white appearance because colour saturation is removed.

Q4: Why might a custom filter begin with convert("RGB")?

So every pixel can be handled consistently as three values: r, g, b.

Q5: What does a threshold filter do?

It compares pixel intensity with a threshold and assigns one result below the threshold and another result above it — commonly black or white.

Q6: What is the negative of (100, 150, 200)?

(155, 105, 55).

Q7: Why is clamping necessary in custom colour filters?

Because standard RGB channel values must remain between 0 and 255.

PIL Image Filters

Learning Objectives

Lesson Presentation

Conceptual Anchor

A Filter Is a Transformation

Built-in Filters — Visual Reference

Original

Gaussian Blur

Sharpen

Find Edges

Emboss

Smooth

Interactive Filter Lab

Using ImageFilter

1 Apply and Save a Built-in Filter

2 Chain Filters

Order Matters

Image Enhancement

Try the factors

3 Increase Brightness

Why Do Blur and Edge Filters Look at Neighbours?

3 × 3 Neighbourhood

Conceptual difference

Custom Filters — Pixel Mathematics

Input Pixel

Formula

Worked Custom Filters

4 Manual Grayscale

5 Negative

6 Sepia

7 Threshold

8 Warm Filter

9 Keep Only the Red Channel

Keep RGB Values in Range

10 Clamp a Calculated Value

A Small Helper Function

Further Viewing

Blurring, Sharpening and Smoothing with Pillow

Common Pitfalls

1. filter() Returns a New Image

2. Unexpected Image Mode

3. Values Outside 0–255

4. Slow Pixel-by-Pixel Code

5. Overwriting Your Original File

Practice Tasks

Self-Check Quiz