Images

Chris Piech and Mehran Sahami

CS106A, Stanford University

Piech + Sahami, CS106A, Stanford University

 Global Variables: Bad Style
# Constant – visible to all functions
NUM_DAYS_IN_WEEK = 7

# Global variable – visible to all functions

balance = 0
Different variables with the same name!
Super confusing!
def main():
balance = int(input("Initial balance: "))
while True:
amount = int(input("Deposit (0 to quit): "))
if amount == 0:
break • Also, really BAD style
deposit(amount)
– So bad, that Python won't even let you do
it unless you basically add a command
that says "I want to have bad style"
def deposit(amount):
balance += amount – I'm not going to show you that command
in Python
– But, if you know it already, DON'T use it!
– We're in polite company
 Using Parameters: Good Style

Don't want using your toaster

to impact your refrigerator!

def main():
balance = int(input("Initial balance: "))
while True:
amount = int(input("Deposit (0 to quit): "))
if amount == 0:
break
balance = deposit(balance, amount)

Encapsulation Principle:
def deposit(balance, amount): Data used by a function
balance += amount
return balance
should be a parameter or
encapsulated in function
 Learning Goals
1. Understanding how images are represented
2. Learning about the SimpleImage library
3. Writing code that can manipulate images

Piech + Sahami, CS106A, Stanford University

 Images

Piech + Sahami, CS106A, Stanford University

 What is an Image?
• Image made of square pixels
– Example: flower.png
• Each pixel has x and y coordinates in the image
– The origin (0, 0) is at the upper-left corner
– y increases going down, x increases going right
• Each pixel has single color encoded as 3 RGB values
– R = red; G = green; B = blue (0,0) x
– Each value represents brightness for
that color (red, green, or blue)
– Can set RGB values to make any color!

y
Piech + Sahami, CS106A, Stanford University
Pixels in an Image Close-Up

Piech + Sahami, CS106A, Stanford University

Working with Images:
Pillow and the
SimpleImage library

Piech + Sahami, CS106A, Stanford University

 Installing Pillow
• Pillow is a version of the Python Imaging Library (PIL)
– Nick Parlante built SimpleImage library using Pillow
– You'll be using SimpleImage in this class
– So, you need to install Pillow first
• To install Pillow, open PyCharm Terminal tab and type
(note the capital P in Pillow):
– On a PC: py -m pip install Pillow
– On a Mac: python3 -m pip install Pillow
– Will see something like:
...bunch of stuff...
Successfully installed Pillow-7.1.1
• Handout #8: Image Reference Guide contains more
information
Piech + Sahami, CS106A, Stanford University
 Using SimpleImage Library
• In folders for assignment or lecture on images, there is
a file simpleimage.py
– This is the SimpleImage library
• To use the SimpleImage library in your code, include at
the top of your program file:
from simpleimage import SimpleImage

• This is importing the SimpleImage module, so that it is

accessible in the code you write
– Similar to when you used import random to use random
number generator library

Piech + Sahami, CS106A, Stanford University

Functions in SimpleImage Library
• Create a SimpleImage object by reading an image from
file (jpg, png, gif, etc.) and store it in a variable.
– Note: each SimpleImage object is made up of Pixel objects
my_image = SimpleImage(filename)

• Show the image on your computer.

my_image.show()

• We can manipulate an image by changing its pixels

• We can also create new images and set its pixels

Piech + Sahami, CS106A, Stanford University

 Accessing Pixels in an Image
• We can use a new kind of loop called a "for-each" loop
• Recall basic for loop (using range):
for i in range(num):
# i will go from 0 to num - 1
do_something()
• For-each loop:
for item in collection:
# Do something with item

• For-each loop with image:

image = SimpleImage("flower.jpg")
for pixel in image:
# Do something with pixel
Piech + Sahami, CS106A, Stanford University
 For-Each Loop Over Pixels

image = SimpleImage("flower.jpg")
for pixel in image:
# Body of loop
This code gets
# Do something with pixel repeated once for
each pixel in image

• Like variable i in for loop using range(),

pixel is a variable that gets updated with
each loop iteration.
• pixel gets assigned to each pixel object in
the image in turn.

Piech + Sahami, CS106A, Stanford University

 Properties of Images and Pixels
• Each SimpleImage image has properties you can access:
– Can get the width and height of image (values are in pixels)
image.width, image.height

• Each pixel in an image also has properties:

– Can get x, y coordinates of a pixel in an image
pixel.x , pixel.y
– Can get RGB values of a pixel
pixel.red, pixel.green, pixel.blue
• These are just integers between 0 and 255
• Higher R, G, or B values means more of that color in pixel
– Can also set pixel RGB values in an image to change it!
Piech + Sahami, CS106A, Stanford University
 Example: A Darker Image

def darker(filename):
"""
Reads image from file specified by filename.
Makes image darker by halving red, green, blue values.
Returns the darker version of image.
"""
# Demonstrate looping over all the pixels of an image,
# changing each pixel to be half its original intensity.
image = SimpleImage(filename)
for pixel in image:
pixel.red = pixel.red // 2
pixel.green = pixel.green // 2
pixel.blue = pixel.blue // 2
return image

Piech + Sahami, CS106A, Stanford University

 Example: Get Red Channel

def red_channel(filename):
"""
Reads image from file specified by filename.
Changes the image as follows:
For every pixel, set green and blue values to 0
yielding the red channel.
Return the changed image.
"""
image = SimpleImage(filename)
for pixel in image:
pixel.green = 0
pixel.blue = 0
return image

Piech + Sahami, CS106A, Stanford University

Let's take it out for a spin!
imageexamples.py

Piech + Sahami, CS106A, Stanford University

 Greenscreening

Piech + Sahami, CS106A, Stanford University

 What is Greenscreening?
• Like the movies (and Zoom backgrounds)
– Have original image with areas that are "sufficiently green."
– Replace "green" pixels with pixels from corresponding x, y locations
in another image

Piech + Sahami, CS106A, Stanford University

 What is Greenscreening?
• Like the movies (and Zoom backgrounds)
– Have original image with areas that are "sufficiently green."
– Replace "green" pixels with pixels from corresponding x, y locations
in another image
INTENSITY_THRESHOLD = 1.6
def greenscreen(main_filename, back_filename):
image = SimpleImage(main_filename)
back = SimpleImage(back_filename)

Piech + Sahami, CS106A, Stanford University

 What is Greenscreening?
• Like the movies (and Zoom backgrounds)
– Have original image with areas that are "sufficiently green."
– Replace "green" pixels with pixels from corresponding x, y locations
in another image
INTENSITY_THRESHOLD = 1.6
def greenscreen(main_filename, back_filename):
image = SimpleImage(main_filename)
back = SimpleImage(back_filename)
for pixel in image:

Piech + Sahami, CS106A, Stanford University

 What is Greenscreening?
• Like the movies (and Zoom backgrounds)
– Have original image with areas that are "sufficiently green."
– Replace "green" pixels with pixels from corresponding x, y locations
in another image
INTENSITY_THRESHOLD = 1.6
def greenscreen(main_filename, back_filename):
image = SimpleImage(main_filename)
back = SimpleImage(back_filename)
for pixel in image:
average = (pixel.red + pixel.green + pixel.blue) // 3
# See if this pixel is "sufficiently" green
if pixel.green >= average * INTENSITY_THRESHOLD:

Piech + Sahami, CS106A, Stanford University

 What is Greenscreening?
• Like the movies (and Zoom backgrounds)
– Have original image with areas that are "sufficiently green."
– Replace "green" pixels with pixels from corresponding x, y locations
in another image
INTENSITY_THRESHOLD = 1.6
def greenscreen(main_filename, back_filename):
image = SimpleImage(main_filename)
back = SimpleImage(back_filename)
for pixel in image:
average = (pixel.red + pixel.green + pixel.blue) // 3
# See if this pixel is "sufficiently" green
if pixel.green >= average * INTENSITY_THRESHOLD:
# If so, overwrite pixel in original image with
# corresponding pixel from the back image.
x = pixel.x
y = pixel.y
image.set_pixel(x, y, back.get_pixel(x, y))
return image Piech + Sahami, CS106A, Stanford University
 Let's try it!
(But using red instead of green)

Piech + Sahami, CS106A, Stanford University

Mirroring an image

Piech + Sahami, CS106A, Stanford University

Piech + Sahami, CS106A, Stanford University
 Nested Loops
image = SimpleImage(filename)
width = image.width
height = image.height

for y in range(height):
for x in range(width):
pixel = image.get_pixel(x, y)
# do something with pixel

y (height) x (width)

Piech + Sahami, CS106A, Stanford University

 Mirroring an Image
def mirror_image(filename):
image = SimpleImage(filename)
width = image.width
height = image.height

# Create new image to contain mirror reflection

mirror = SimpleImage.blank(width * 2, height)

for y in range(height):
for x in range(width):
pixel = image.get_pixel(x, y)
mirror.set_pixel(x, y, pixel)
mirror.set_pixel((width * 2) - (x + 1), y, pixel)
return mirror

Piech + Sahami, CS106A, Stanford University

I wanna see it!

Piech + Sahami, CS106A, Stanford University

 What's The Difference?

def darker(filename): def darker(filename):

img = SimpleImage(filename)
img = SimpleImage(filename)
for y in range(img.height):
for px in img: for x in range(img.width):
px.red = px.red // 2 px = img.get_pixel(x, y)
px.green = px.green // 2 px.red = px.red // 2
px.blue = px.blue // 2 px.green = px.green // 2
return img px.blue = px.blue // 2
return img

Nothing!
We only want to use nested for loops if
we care about x and y.
(Needed that for mirroring image.)
Piech + Sahami, CS106A, Stanford University
 Learning Goals
1. Understanding how images are represented
2. Learning about the SimpleImage library
3. Writing code that can manipulate images

Piech + Sahami, CS106A, Stanford University

Piech + Sahami, CS106A, Stanford University
Piech + Sahami, CS106A, Stanford University