Image Compression

Content

 Redundancies
 Coding
 Interpixel
 Psychovisual
 Fidelity Criteria
 Image Compression models
 Error free compression
 Lossy compression
 Image compression standards
 Compression..??
 What is compression.?
 A technique by which the data is stored as information.
 Representation of data in such a way that it makes the
use of less resources to store and transmit it.
 How it is done..??
 By reducing the redundancy present in the data.
 Redundancy = Data – Information
 Redundancy = Irrelevant or symmetric data.
 R= 1- 1/C
 Where
 c= b/b’
 R = relative redundancy
 b= bits used to represent data
 b’ = bits used to represent the information from that data
Redundancies

 Coding Redundancies
 Interpixel Redundancies
 Psychovisual Redundancies
 Coding Redundancies
 Code: system of letters, numbers, bits (symbols)
 Codeword: Meaningful sequence of symbols
 Code Length: Number of symbols used to form codeword
 Often Coding techniques makes the use of more number of bits than
that required to store the intensity.
 Example
 In 8 bit gray scale representation intensity level ‘6’ is represented by
‘00000110’
 where only 3 last bits are significant and rest 5 are redundant.
 In this example
 Code = {1,0}, Codeword = 00000110 and Code Length = 8
 Other way to exploit redundancy
 Suppose we are making use of 8 bits to represent the gray levels in the
image.
 But suppose image is having only ’64’ different intensity levels in it out
of possible ‘256’ gray levels.
 Now if we assign new code to those ’64’ intensities, ‘6’ bits are sufficient
to store the pixel information
Interpixel Redundancy (based
on planes)
 Spetial (Intra plane) Redundancy:
 The correlated pixels in a still image

 Temporal (Interplane) Redundancy:

 The correlated pixels in two consecutive
frames (two still images)
Psychovisual Redundancies

 The information coded, that is not

interpreted by human visual system.
 The information coded, that is not
detected by human visual system.
 Detected = Interpreted..??
Fidelity Criteria
 How to specify Information loss.?
 Objective fidelity (mathematical)
 Subjective fidelity (human opinion)
 Objective Fidelity
 Let f(x,y) is original image and f’(x,y) be the
approximation of f(x,y)
 Error e(x,y) = f’(x,y) – f(x,y)
 So the total error between original and
decompressed image
 ∑∑[f’(x,y) – f(x,y)]
 Where summation is taken along the rows and
columns
Fidelity Criteria (cntd…)

 Root Mean Square Error

 Where MSE is mean square error.

 MSE =1/MN ( ∑∑[f’(x,y) – f(x,y)] )2

 Where summation is taken along the rows
(M) and columns (N)
Fidelity Criteria (cntd…)

 Signal to Noise Ratio (SNR)

 ∑∑f’(x,y)2 / ∑∑[f’(x,y) – f(x,y)] )2
 Where summation is taken along the rows
(M) and columns (N)

 i.e. Approximated Signal / Square Error

Fidelity Criteria (cntd…)
 Subjective Fidelity
 Human opinion about the perceived image
 Excellent (much better)
 Fine (better)
 Passable (slightly better)
 Marginal (slightly worse)
 Inferior (worse)
 Unusable (much worse)

 These comments can be ranged from -3 to 3

numeric from ‘much better’ to ‘much ‘worse’
respectively
Image Compression Model

 General Compression Model

 Generic Image Compression Model

 Special about Image Compression
 Compressing an image is significantly different than compressing raw
binary data.
 Of course, general purpose compression programs can be used to
compress images, but the result is less than optimal.
 This is because images have certain statistical properties which can be
exploited by encoders specifically designed for them.
 Also, some of the finer details in the image can be sacrificed for the
sake of saving a little more bandwidth or storage space.
 This also means that lossy compression techniques can be used in this
area. 
 Lossless compression involves with compressing data which, when
decompressed, will be an exact replica of the original data.
 This is the case when binary data such as executables, documents etc.
are compressed.
 On the other hand, images (and music too) need not be reproduced
'exactly'.
 An approximation of the original image is enough for most purposes, as
long as the error between the original and the compressed image is
tolerable.
 Image Compression Model (cntd..)

 Mapper : Reduces inter-pixel redundancy

 Run length coding: reversible
 Calculation of DCT: reversible
 Quantizer: reduces psyco-visual redundancy
 Reduction of grey scales: not reversible
 Removal of high frequency content: not reversible
 Reduces the accuracy of Mapper’s output
 Symbol Enoder:reduces coding redundancy
 Generates the fixed length or variable length symbols
to represent the output of quantizer.
 Variable length coding: reversible
 Channel transmits the compressed data.
Error Free Compression
Variable length coding
 Huffman coding
 Most probable symbol is assigned the shortest code
word

 Some Properties of Huffman Code

 Instantaneous:
 do not refer|ence s{uz cceed}ing symbols
 Uniquely decodable
 symbols can be decoded in only one way
 Block code
 for each symbol: only one code
Example
 Symbol Frequency
 A 24
 B 12
 C 10
 D 08
 E 08
 Total 186 bit (with 3 bit per code word)
 Steps
 The two rarest symbols 'E' and 'D' are connected first,
followed by 'C' and 'D'.
 The new parent nodes have the frequency 16 and 22
respectively and are brought together in the next step.
 The resulting node and the remaining symbol 'A' are
subordinated to the root node that is created in a final
step.
Example Continued
Example Continued

S F C CL TL
 A 24 0 1 24
 B 12 100 3 36
 C 10 101 3 30
D 8 110 3 24
E 8 111 3 24
 -------------------------------------
 186 138 bits
 Lavg = (0.4)(1) + (0.3)(2) + (0.1)(3) + (0.1)(4) + (0.06)(5) +
(0.04)(5) = 2.2 bits/pixel
Arithmetic Coding

 Huffman uses a static table for the whole

coding process, so it is rather fast, but
does not produce an efficient
compression ratio.
 Arithmetic coding, on the other hand,
has different features.
 It can generate a high compression ratio (3
bits per character on avarage), but all the
complex calculation takes much more time,
resulting in a slower implementation.
 AAA: 0, 0.0996
AA: 0, 0.166 AAB: 0.0996, 0.1328
AAC: 0.1328, 0.166
ABA: 0.166, 0.1869375
: 0, 0.333 AB: 0.166, 0.24975 ABB: 0.1869375, 0.2288125
ABC: 0.2288125, 0.24975
ACA: 0.24975, 0.2705625
AC: 0.24975, 0.333 ACB: 0.2705625, 0.291375
ACC: 0.291375, 0.333
BAA: 0.333, 0.34965
BA: 0.333, 0.41625 BAB: 0.34965, 0.3996
BAC: 0.3996, 0.41625
BBA: 0.41625, 0.457875
B: 0.333, 0.666 BB: 0.41625, 0.58275 BBB: 0.457875, 0.541125
BBC: 0.541125, 0.58275
BCA: 0.58275, 0.5994
BC: 0.58275, 0.666 BCB: 0.5994, 0.6327
BCC: 0.6327, 0.666
CAA: 0.666, 0.68265
CA: 0.666, 0.74925 CAB: 0.68265, 0.6993
CAC: 0.6993, 0.74925
CBA: 0.74925, 0.7700625
C: 0.666, 1.0 CB: 0.74925, 0.8325 CBB: 0.7700625, 0.790875
CBC: 0.790875, 0.8325
CCA: 0.8325, 0.874375
CC: 0.8325, 1.0 CCB: 0.874375, 0.91625
Initial:
D
i
v
i
d
e
a
r
e
a
i
n
2nd
t
1st Iteration: Iteratio
o
After n: 3rd Iteration: After
5 4th Iteration: After
A After ABA 5th Iteration: After
( ABAC
Divide AB Divide ABACU
# Divide
previo Divide previous Divide previous
o previous area
us area previo area by 8 (# area by 10 (#
f by 9 (#
by 6 (# us area unique unique characters
u unique
unique by 7 (# characters +5)
n characters +4)
charact unique +3)
i
ers +1) charact
q
ers +2)
u
e
c
h
Comparison
 Lampel-Zev-Welch
 Huffman coding and Arithmetic coding
 Coding redundancy
 LZW emphasizes on Spatial Redundancy
 Error free compression technique that makes the use of
fixed length codes.
 It assigns fixed length-length to variable length code words.
 It is a patent technique 
 Technique is used in GIF, TIFF and PDF
 So PNG was designed to overcome the difficulties of patent .
 LZW gives the better compression ratio than that of HC or
AC as it exploits both coding as well as spetial
redundancies.
How LZW works.?
 A dictionary is constructed which contains
 Source symbols to be encoded.
 For 8 bit image, intensities [0,255] are put in
dictionary
 Encoder examines the pixel intensities
 Intensity that are not in the dictionary are put in
next unused (algorithmically determined)
locations.
 Suppose we encountered two consecutive
pixels of value 255 then a new code word ‘256’
is used to represent it.
 Number of bits saved = (8+8) – 9 = 7 bits
 Introduction to LZW (cont'd)
 Codes 0-255 in the code table are always assigned to represent single
bytes from the input file.

 When encoding begins the code table contains only the first 256
entries, with the remainder of the table being blanks.

 Compression is achieved by using codes 256 through 4095 to represent

sequences of bytes.

 As the encoding continues, LZW identifies repeated sequences in the

data, and adds them to the code table.

 Decoding is achieved by taking each code from the compressed file,

and translating it through the code table to find what character or
characters it represents.
Example 1: Compression using LZW

Example 1: Use the LZW algorithm to compress the string

BABAABAAA
 Example 1: LZW Compression Step 1

BABAABAAA P=A
STRING TABLE
C=empty
ENCODER OUTPUT
string codeword representing output code
BA 256 B 66
 Example 1: LZW Compression Step 2

BABAABAAA P=B
STRING TABLE
C=empty
ENCODER OUTPUT
string codeword representing output code
BA 256 B 66
AB 257 A 65
 Example 1: LZW Compression Step 3

BABAABAAA P=A
STRING TABLE
C=empty
ENCODER OUTPUT
string codeword representing output code
BA 256 B 66
AB 257 A 65
BAA 258 BA 256
 Example 1: LZW Compression Step 4

BABAABAAA P=A
STRING TABLE
C=empty
ENCODER OUTPUT
string codeword representing output code
BA 256 B 66
AB 257 A 65
BAA 258 BA 256
ABA 259 AB 257
 Example 1: LZW Compression Step 5

BABAABAAA P=A
STRING TABLE
C=A
ENCODER OUTPUT
string codeword representing output code
BA 256 B 66
AB 257 A 65
BAA 258 BA 256
ABA 259 AB 257
AA 260 A 65
 Example 1: LZW Compression Step 6

BABAABAAA P=AA
STRING TABLE
C=empty
ENCODER OUTPUT
string codeword representing output code
BA 256 B 66
AB 257 A 65
BAA 258 BA 256
ABA 259 AB 257
AA 260 A 65
AA 260
 LZW Decompression

 The LZW decompressor creates the same string table during

decompression.

 It starts with the first 256 table entries initialized to single characters.

 The string table is updated for each character in the input stream,
except the first one.

 Decoding achieved by reading codes and translating them through the

code table being built.
 LZW Decompression Algorithm

1 Initialize table with single character strings

2 OLD = first input code
3 output translation of OLD
4 WHILE not end of input stream
5 NEW = next input code
6  IF NEW is not in the string table
7 S = translation of OLD
8   S=S+C
9 ELSE
10  S = translation of NEW
11 output S
12   C = first character of S
13   OLD + C to the string table
14 OLD = NEW
15 END WHILE
 Run length Encoding

 It specifies two values

 New intensity value
 Number of consecutive pixels having that intensity.
 The technique was explored for two dimensions
 It became standard in facsimile (FAX)
 It exploits the ------------ redundancy..??
 It exploits spetial redundancy by grouping
consecutive pixels having same intensity value.
RLE Example
Standards that use RLE

 CCITT
 JPEG
 M-JPEG
 BMP
 Symbol Based Coding
 It exploits coding as well as spetial
redundancies
 It assign a unique value to each symbol.
 For compression it stores location and
corresponding code of that symbol.
 Bit Plane Coding
 First make the use of bit slice of the image
 Then use one of the available techniques to compress the file.
 Generally
 Image is broken up into bit planes and then RLE is applied to each
plane.
 Let the image is represented as
 am-12m-1 + am-2 2m-2 + a1 21 + a0 20
 So here we will get ‘m’ planes (0,1,2...m-1)
 Problem:
 Small change in intensities can cause lots of variation in code i.e. 127
(01111111) and 128(10000000)
 Solution
 Make the use of gray code
 gi = ai EXOR ai+1 where 0<=i<=m-2 and gm-1 = am-1
 Block Transform Coding

I/P Block Symbol Compressed

FT Quantizer Image
Preparation Encoder

Merge all Inverse Symbol

blocks Transform Decoder
 Block Preparation

 Y = 0.30R+0.59G+0.11B
 I = 0.60R-0.28G-0.32B
Q = 0.21R-0.52G+0.31B
 4800 blocks
Block ‘Y’ matrix to DCT
Computation of the quantized DCT coefficients
Transmission order
Image Compression
Standards
 JPEG
 CCITT
 BMP