Inft1004_Lec5_MorePictureTechniques

7/27/2019 Inft1004_Lec5_MorePictureTechniques

1/49

INFT1004Visual Programming

Lecture 5

More Image Manipulation

(Guzdial & Ericson chapter 5)


2/49

3

INFT1004 - SEMESTER 1 - 2012

Week 1

Week 2

Week 3

Week 4

Week 10

Week 11

Week 9

Week 12

Mar 4

Mar 11

Mar 18

Mar 25

May 6

May 13

May 20

May 27

Assignment due3:00 Tuesday May 21

Programs, Arrays and Iteration

Working with x and y coordinates

Selection

Web, Representations, Steganography

Lists, Files and Modules

Turtles and Other Classes

Introduction

Week 13 Jun 3

Revision and Look Ahead

No formal classes

Mid Year Examination Period - MUST be available normal & supplementary period

Recess Apr 1 - Apr 7 Mid Semester Recess Period

Practical Test 2in Lab class

Week 8 Apr 29

More Sound and ArraysWeek 7 Apr 22

Program Design and Strings

Lecture Topics and Lab topics are the same for each week

LECTURE TOPICS

Week 5 Apr 8 Practical Test 1in Lab class

More Picture Techniques

Week 6 Apr 15 Sound and Arrays


3/49

4

INFT1004 - SEMESTER 1 - 2012

Week 1

Week 2

Week 3

Week 4

Week 10

Week 11

Week 9

Week 12

Mar 4

Mar 11

Mar 18

Mar 25

May 6

May 13

May 20

May 27

Assignment due3:00 Tuesday May 21

Programs, Arrays and Iteration

Working with x and y coordinates

Selection

Web, Representations, Steganography

Lists, Files and Modules

Turtles and Other Classes

Introduction

Week 13 Jun 3

Revision and Look Ahead

No formal classes

Mid Year Examination Period - MUST be available normal & supplementary period

Recess Apr 1 - Apr 7 Mid Semester Recess Period

Practical Test 2in Lab class

Week 8 Apr 29

More Sound and ArraysWeek 7 Apr 22

Program Design and Strings

LECTURE TOPICS

Week 5 Apr 8 Practical Test 1in Lab class

More Picture Techniques

Week 6 Apr 15 Sound and Arrays

You must attend the lab you are

registered in or you will not be

able to do the practical test


4/49

5

Revision Simple & Object Types

With simple types (eg int, float, string)

Assignment gives the value of the expression on

the right to the variable on the left

number = 8

sum = 10 + number


5/49

6

Revision Simple & Object Types

With object types

Assignment makes the variable on the left a name

(or another name) for the object on the right

myFile = newFile

newFile

myFile


6/49

7

Revision - Functions

Programs are made up of one of more functions

If we want a function to take arguments, we includecorrespondingparameters when defining it

Whenever we find the same piece of code appearingseveral times in a program, we should extract it, define it as

a function, and replace the multiple occurrences withmultiple calls to the function

If there are variations between the occurrences, we useparameters and arguments to deal with the variations


7/49

8

Revision - sequence, selection

and iteration

The statements in a function or a loop will be

executed in the order theyre written

When working out what the code does, be sure tonote the effect of each statement before going on

to consider the next

sequence


8/49

9

Within a sequence of statements, a loop iterates

(repeats) the statements in its body, as manytimes as specified

sequence loop


and iteration


9/49

10

In a selection statement, the program chooses

which bits of code to execute depending on thevalue of a boolean condition


and iteration

sequence loop

selection?


10/49

11

Selection is done with the ifstatement

The body of an ifstatementis executed if the

condition is true

If theres an else part, the body of the else is

executed if the condition is false

Revision forms of selection


11/49

12

If we want to test multiple options, we generally useelif, an abbreviation of else if

The first body of an elifthat has a true condition is

executed (all preceding conditions must be false)

After any body (if, elif, else) is executed, the rest ofthe ifsequence is skipped

Revision forms of selection


12/49

13

Revision - Comments

Your programs must include three kinds of comments:

1.A comment at the start of every program (collection offunctions), saying who wrote it, and when, and why

2.A comment at the start of every function, explaining brieflywhat it does3.A comment with every bit of code that another programmermight find easier to understand if its explained

Do not write comments to explain what will be

obvious to a reasonable programmer!


13/49

14

Oversight makeEmptyPicture()

Earlier lectures overlooked something thats pretty

important in image manipulation

Weve always manipulated an image itself,

changing the original

Sometimes one might want to keep the original and

work with a copy


14/49

15


Once you have the height and width of a picture,

makeEmptyPicture(width, height)

can give you a new picture of the required size

You can then copy (the colours of) the pixels from

the original to the copy, and then manipulate the

copy without altering the original


15/49

16

Oversight makeEmptyPicture()width

height

makeNewPicture(width, height)


16/49

17


So you can use this to:

1)create a new picture and2)

copy (the colours of) the pixels from the originalto the copy

3)manipulate the copy without altering the original


17/49

18

More Image Manipulation

Theres very little new programming in this lecture

Instead we revise the programming weve done by

seeing some more approaches to manipulating

images

Well see how to blend images, subtract

background, use chromakey, and draw and write on

images


18/49

19

Blending images

To get the effect of transparency, looking throughone somewhat transparent image at anotherimage, all we have to do is average the colours of

the corresponding pixels.


19/49

20

Blending images

To overlay picture 1 and picture 2, we change the pixelsof picture 1 to be the averages of the pixels of the two

pictures

If we do a 50/50 average we get half of each picture;

different proportions give different apparent transparency


20/49

21

Blending images

To overlay picture 1 and picture 2, we change the pixelsof picture 1 to be the averages of the pixels of the two

pictures

0 1 2 3 4 5

0

1

2

3

4

5

0 1 2 3 4 5

0

1

2

3

4

5

0 1 2 3 4 5

0

1

2

3

4

5

newPic1Pixel = (oldPic1Pixel*0.5) + (pic2Pixel*0.5)

newPic1Pixel pic2PixeloldPic1Pixel

If we do a 50/50 average we get half of each picture;

different proportions give different apparent transparency


21/49

22

Transparent overlay

Lets overlay a picture of a butterfly on apicture of a bridge

butterfly1.jpg

422

497

bridge.jpg

640

480


22/49

23

Transparent overlayWe need to specify at which coordinates of picture 1

(bridge) to start overlaying picture 2 (butterfly)

We need to specify how transparent to make picture 2

xstart

ystart 50%

transparent


23/49

24

Conditions

70%

transparent

40%

transparent


24/49

25

Conditions

If we want picture 2 to be 70%

transparent, we blend the colour

channels to be 70% of the pixel frompicture 1 and 30% from picture 2

70%

transparent

(oldPic1Pixel*0.7) + (pic2Pixel*0.3)

70%

30%

picture 1

picture 2


25/49

26

ConditionsIf we want picture 2 to be 40%

transparent, we blend the colourchannels to be 40% of the pixel frompicture 1 and 60% from picture 2

40%

transparent (oldPic1Pixel*0.4) + (pic2Pixel*0.6)

40%

60%

picture 1

picture 2


26/49

27

Conditions

(oldPic1Pixel*(n/100)) + (pic2Pixel*(100-n)/100)

n%

transparent

n%

(100 - n)%

In general, if we want picture 2 to be n% transparent,

we blend the colour channels to be n% of the pixel

from picture 1 and (100 n)% from picture 2

picture 1

picture 2


27/49

28

Example

(oldPic1Pixel*(80/100)) + (pic2Pixel*(100-80)/100)

80%

transparent

n%

(100 - 80)%

e.g. if we want picture 2 to be 80% transparent,

we blend the colour channels to be 80% of the pixel

from picture 1 and (100 80)% from picture 2

picture 1

picture 2

(oldPic1Pixel* 0.8)) + (pic2Pixel* 0.2)


28/49

29

Example

See transparency() in lecture5demo.pydef transparency(pic1, pic2, xstart, ystart, proportion):

# Overlay pic2 on pic1, starting at coordinates xstart and ystart of pic1,

# with proportion determining the transparency (0-100) of the overlay

for x in range(0, getWidth(pic2)):

for y in range(0, getHeight(pic2)):

# x and y are coordinates in pic2; add xstart and ystart to get the matching coords in pic1

pix1 = getPixel(pic1, xstart + x, ystart + y)

pix2 = getPixel(pic2, x, y)

# Blend the channels: proportion% of pic1 and (100-proportion)% of pic2

newRed = getRed(pix1) * proportion / 100 + getRed(pix2) * (100 - proportion) / 100

newGreen = getGreen(pix1) * proportion / 100 + getGreen(pix2) * (100 - proportion) / 100

newBlue = getBlue(pix1) * proportion / 100 + getBlue(pix2) * (100 - proportion) / 100

newColour = makeColor(newRed, newGreen, newBlue)

# And apply this new blended colour to the pixel of pic1

setColor(pix1, newColour)

repaint(pic1)


29/49

30

Fading

The book blends one picture into another by doinga section of picture 1 alone, a section of 50/50

transparency, and a section of picture 2 alone

This is worth examining to see how they manage

the coordinates of picture 1, picture 2, and the new

picture

(They paint to a new picture, whereas were

altering picture 1)


30/49

31

Fading

But were going to do something quite different:

A full fade, from 0 transparency at the top of picture 2

to 100% transparency at the bottom

0%

100%

transparency

picture 2


31/49

32

Fading

y coordinate

Every y value needs its own transparency level

0%

100%

0

getHeight(picture2)

transparency


32/49

33

Fading

y coordinate

How do the transparency levels relate to the coordinates?

proportion = 100 * y / getHeight(pic)

0%

100%

0

getHeight(picture2)

transparency


33/49

34

Varying transparency levels

How do we work that out?

By problem solving looking at what were given

and using it to work out what we want

You cant avoid problem solving in programming!

See fadeIn() in lecture5demo.py


34/49

Subtracting Background

When we overlaid the picture of the butterfly on the bridge,the picture of the butterfly was a rectangle that included

background

35


35/49


When we overlaid the picture of the butterfly on the bridge,the picture of the butterfly was a rectangle that included

background

It might be good to get rid of the

background, and only overlay thebutterfly itself (the foreground)

That is, we want to subtract the

background from the picture

To do this, we need to know whichpixels are background and which

are foreground36


36/49


If we have one picture of the background alone,and one of the background with foreground . . .

We can compare the two, and choose only the

pixels that are different

background background + foreground 37


37/49


That way, we get just the foreground pixels.

38background + foreground foreground


38/49


Which we can overlay on a third picture, (a new

background)

39foreground + new backgroundforeground


39/49

40

Where it falls over

Ive used a gif butterfly with transparent background to

illustrate overlaying background and foreground.

If you use this in the real world say you take a picture of

something in front of a background and then a separate

picture of the background - you may encounter a couple of

problems.

backgroundforeground + background


40/49

41

Where it falls over

Shadow is really background, but its different from thebackground picture, so it registers as foreground

The problem with shadow can be fixed with direct lighting

shadow


41/49

42

Where it falls over

Another problem is that any colour in the foreground that is

very similar to the background registers as background

The problem with similarity between foreground and

background can be fixed only if we can ensure there is no

such similarity

some foreground colours

match background colours


42/49

43

Chromakey

Chromakey is a technique for subtracting background that

relies on the background being as close as possible to asingle colour . . .

And a colour that doesnt occur very often in the

foregrounds that we tend to work with

Why dont weather forecasters wear blue shirts?


43/49

44

Chromakey

Because theyre filmed

against a blue background,

and blue is then eliminatedfrom the picture in a

background subtraction

(green is also used a lot neither blue or green occur

in skin colours but watch what you wear)


44/49

45

Chromakey

If the foreground is photographed against a blue

background . . .

The function becomes simpler, because its always

comparing with the same colour . . .

And the result is a lot cleaner


45/49

46

Drawing on picturesLike most programming languages, JES lets us add

certain features directly to a picture

addRect(pic, x1, y1, w, h)

All of these can take an additional argument, colour

addLine(pic, x1, x2, y1, y2)

addOvalFilled(pic, x1, y1, w, h)

addOval(pic, x1, y1, w, h)

addRectFilled(pic, x1, y1, w, h)

a line from (x1, y1) to (x2, y2)

a rectangle of width w andheight h starting at (x1, y1)

a filled rectangle

an oval that fits in a rectangle

of width w and height hstarting at (x1, y1)

a filled oval


46/49

47

Drawing on pictures


47/49

48

Writing on pictures

This, too, can take an additional argument, colour

Does the top of the string or the bottom of the string line upwith the vertical coordinate y?

How would you find out?

To alter the font or size see Ch16 page 398 and look atthe JES Pictures functionaddTextWithStyle()

addText(pic, x, y, string)the string (of text) starting

at location (x, y)


48/49

49

Writing on pictures


49/49

50

Moving on from picturesThats all for pictures for the time being. Next week well

start working with sound files

The MediaTools program will be very useful for that, so

if you havent yet installed it, now would be a good time

50

Inft1004_Lec5_MorePictureTechniques

Documents