1

CHAPTER 1

INTRODUCTION

Information sharing is gaining importance, especially after 9/11, in

order to prevent terrorism attacks. Governments refine Intelligence Reform

and Prevention of Terrorism policies continuously to provide a better

information sharing environment across the globe. Traditional information

sharing happens as one-to-one exchange of data between sender and receiver

using open and proprietary protocols, messages and file formats. During late

1970s, modern information sharing methods started using Electronic Data

Interchange (EDI) standard for data exchange. From then on, EDI has become

the de-facto standard till date for data exchange. Technologies such as blogs,

Wikis and social networks continuously evolve to fit into one of the following

information sharing design patterns: one-to-one, one-to-many, many-to-many,

and many-to-one. As technologies evolve, information sharing platforms need

to provide and facilitate controlled vocabularies, data harmonization, privacy

and security policies.

End users/customers prefer secured communication during

information sharing without the apprehension of third parties or an attacker

intercepting what is being shared. In information sharing, huge volume of

data that need to be transferred becomes the bottleneck in providing secured

communication. Secured communication is not just any communication that

employs encryption, but it should posses the ability to process rapidly

growing information. It is imperative to provide efficient, secured solutions

depending on the demands and needs of end-users during information sharing.

2

There are several security algorithms proposed in the literature for secured

communication. They are expected to provide the following essential

services: Privacy, Authentication, Integrity and Non-Repudiation. These

services are collectively called as PAIN services. Secured communication

becomes an essential phenomenon in various activities done on daily basis

both in governmental, non-governmental and business organizations. Some of

those activities are listed below:

a) Protecting sensitive information related to government

agencies dealing with social security numbers, bank account

numbers, credit card numbers and national top secrets and

policy decisions

b) Secured data links for defense and military related

communication to and from systems setup in unconventional

places such as satellite, airborne, ground and sea-based remote

platforms

c) Secured Intelligence, Surveillance and Reconnaissance (ISR)

systems that are employed in various social, economic, and

commercial agencies

d) Security in strategic and tactical signal intelligence systems

that detect, collect, identify, analyze and disseminate

information. Few examples are weather forecasting, enemy

trespassing activities, identifying wealth and natural resources

in a particular area etc.

e) Protection in sharing patient’s information such as

Electrocardiograms (ECGs), Magnetic Resonance

Imaging (MRIs), and Digital Imaging and Communications in

Medicine (DICOM)

3

f) Secured communications and solutions such as secured

pathways to communicate sensitive messages among clients,

advocates and judges in legal industry

With rapid growth in digital communication, sharing and

transmission of videos and images over the Internet is unsecured due to

eavesdropping. Various cryptography algorithms are employed in order to

provide secured data communication by preventing cipher attacks and

snooping. Cryptography is an ancient art, originated way back in 4000 BC, to

protect information. Over century of years, cryptographic techniques are

evolved, refined and adopted for providing fast and secured communication

(Chung 2005), (Grangetto 2006), (Hyungjin 2007) (Matthias 2008) .

Rest of this chapter presents the timeline of cryptography

algorithms and basic concepts of video encryption. Various types of

cryptography algorithms are explained with suitable diagrams and examples.

This work focuses mainly on several video security mechanisms.

1.1 THE SECURITY MODEL

Pervasive usage of image, audio and video information require

media content protection making security an indispensable, vital factor. In

order to guarantee utmost data security, access rights alone are not adequate.

Data must be protected along the entire route/path during transmission by

employing suitable encryption techniques in order to provide complete

protection from piracy or unauthorized access. Security algorithms provide

mechanisms to solve risks and/or security threats during information sharing

with the following specific requirements (Gary 1999):

4

a) Privacy or Confidentiality: It makes sure that no one, except

the intended receiver, can read the message.

b) Authentication: It is the process of confirming identity of

users to access / modify the message. i.e. it makes sure that the

user is actually who he/she claims to be.

c) Integrity: It guarantees that the content of the message

received is not tampered; it protects the information from

damage or deliberate manipulation and thereby assuring the

receiver that the received message has not been altered in any

way from the original.

d) Non-repudiation: It means to deny something. Non-

repudiation is the ability to ensure that a sender cannot deny

the authenticity of his/her signature or sending of the message.

It provides a mechanism to prove that the sender has really

sent this message.

The above cryptographic services are realized by employing

encryption and authentication primitives, and cryptographic protocols.

Encryption primitives are used to provide confidentiality, whereas,

authentication primitives are used to provide data authentication. Ciphers

created by employing various encryption algorithms are known as encryption

primitives. Hash functions and digital signatures are examples for

authentication primitives.

5

Figure 1.1 Typical Security Model

A typical security model for information sharing is shown in Figure

1.1 (William 2005). The message is to be transferred from one party to

another across shared transmission medium such as Internet. In the model

shown in Figure 1.1, a logical information channel is established by defining a

route through the Internet from source to destination with the use of

communication protocols agreed upon by both the parties. The model shown

in Figure 1.1 has a trusted third party which facilitates secure transmission.

For example, the third party may be responsible for distributing the secret

information to sender and receiver and arbitrating disputes between them. The

general model in Figure 1.1 shows that there are four main tasks in designing

a particular security service (William 2005):

a) Designing an algorithm for converting the plain text to cipher

text. The designed algorithm should not be broken by

opponents. Various encryption algorithms have been

developed for this purpose (Daniel 2007), (Suman 2011),

(Yuefa 2012).

Opponent Intercepting the

Channel

6

b) Generating the secret information, i.e. the cipher-text, using

encryption algorithms.

c) Developing methods to distribute and share the secret

information.

d) Specifying a protocol for secured communication in order to

achieve a specific goal. e.g. (i) protocols for user

authentication and (ii) protocols for key management.

1.2 EVOLUTION AND HISTORY OF CRYPTOGRAPHY

TECHNIQUES

The word Cryptography is coined from the following two Greek

words: krypto=“hidden” and grafo=“to write” i.e. “hidden writing”.

Cryptography is being used since ancient times in order to hide the original

message during data transfer. Cryptography techniques developed so far can

be divided into the following three eras:

a) Techniques developed from ancient civilization to first part of

nineteenth century (4000 BC – 400 AD). These techniques are

simple algorithms designed and implemented manually.

b) Classical techniques that are used in electro-mechanical

machines (1000 AD - 1948). These algorithms were

developed during world war times, especially for secret

communications between Presidents and Heads of allied

nations.

c) Techniques that are developed during modern computing era

i.e. after the release of microprocessors (1949 – till date).

7

These techniques are developed based on solid mathematical

background.

The history and timeline of cryptography algorithms is shown in

Table 1.1. It is interesting to note that beside researchers and academicians,

other communities such as military persons, diarists and diplomatic personals

have also contributed several techniques in cryptography. For brevity, only

limited details such as name(s) of the invention and inventors are given in

Table 1.1. More details on some of the popular inventions are given

subsequently after Table 1.1.

Table 1.1 History and Timeline of Cryptography Techniques

Period/Year Name(s) of Inventors and Description of the Technique

Ancient Cryptography Techniques

Before Christ Era (BC)

Hieroglyphic Symbols used by Egypts, 4000 BC – AD 400, to transfer religious literature and sacred writing

Phaistos Disks made using clay by Greeks, 1800–1600 BC, first movable type printing

Substitution Cipher, Steganography, Caesar Cipher by Romans -600 BC

Spartans Skytale Device by Greeks, 500 BC

Classical

1000 AD Frequency Analysis by Arabs

1467Cipher Disk

Polyalphabetic Cipher by Leon Battista Alberti used in Mechanical Cipher Machines

1585 Vigenere by Bellaso

1795Jefferson Disk by Thomas Jefferson used in World War II by US Navy

8

Table 1.1 History and Timeline of Cryptography Techniques (Contd.)

Period/Year Name(s) of Inventors and Description of the Technique

1932

Enigma used for Military Communications in World War II

SIGABA used in World War II by USA

Typex used in Rotor Machines by British

1940 One Time Pad used in Banking initiated by Frank Miller

1942 SIGSALY used in World War II

Modern Computing

1949 Mathematical Theory of Cryptography published by Shannon

1970 Quantum States by Stephen Wiesner

1973 Feistel Network Block Cipher Design by Horst Feistel

1975 Public-key Cryptography

1976 Key Exchange Algorithms by Diffie-Hellman-Merkel

1977

Data Encryption Standard (DES) by USA used for enciphering PIN numbers and bank transactions

RSA by Ronald Rivest, Adi Shamir, and Leonard Adleman, used for Secured Communication

1982 Feynman ciphers by Richard Feynman

1984

BB84 -First Quantum Cryptography Protocol designed by Charles Bennett and Gilles Brassard

Probabilistic Encryption by Shafi Goldwasser and Silvio Micali

Chaotic Encryption by Matthews

1994Peter Shor Algorithm

Tiny Encryption Algorithm by David J. Wheeler and Roger M. Needham

1995SECMPEG by Jurgen Meyer, Frank Gadegast used for Video Encryption

1996Zig-Zag Permutation Algorithm by Lei B. Y, Lo K. T and Haijun Lei used for Text, Image and Video Encryption

9

Table 1.1 History and Timeline of Cryptography Techniques (Contd.)

Period/Year Name(s) of Inventors and Description of the Technique

1998

Twofish by Bruce Schneier, John Kelsey, Doug Whiting, DavidWagner, Chris Hall, and Niels Ferguson

Quantum Cryptography

Video Encryption Algorithm (VEA) by Changgui Shi and Bharat Bhargava

MPEG Video Encryption Algorithm (MVEA) by Changgui Shi and Bharat Bhargava

1999Real-time Video Encryption Algorithm (RVEA) by ChangguiShi, Wang SY and Bharat Bhargava

2000 Partial Encryption by Howard Cheng and Xiaobo Li

2001 Advanced Encryption Standard (AES)

2003Frequency Domain Scrambling approach by Wenjun Zeng and Shawmin Lei

Selective Encryption by Xiliang Liu and Ahmet M. Eskicioglu

2005 MHT Scheme by Chung Ping Wu and Jay Kuo C. C

2006Wavelet Packet Transform Algorithm by Dominik Engel and Andreas Uhl

2007

PRESENT Algorithm by Andrey Bogdanov, Lars R. Knudsen, Gregor Leander, Christof Paar, Axel Poschmann, Matthew J. Robshaw, Yanick Seurin and C. Vikkelsoe used for VLSI chip design

2009LCASE (Lightweight Cellular Automata-based Symmetric-key Encryption) by Somanath Tripathy and Sukumar Nandi

2010International Data Encryption Algorithm by Rajashekhar Modugu, Yong-Bin Kim and Minsu Choi

2011Humming Bird-2 by Daniel Engels, Markku-Juhani O. Saarinen, Peter Schweitzer and Eric M. Smith

10

It is believed that Egyptians used a symbolic representation called

Hieroglyphic symbol for recording and transferring religious, secret and

mystical knowledge to next generation around 4000 BC. Hieroglyphs are

figurative, stylized symbols representing real or illusional objects. Though

numerous failed attempts were made to decipher the hidden meaning of

hieroglyphic symbols, real breakthrough came during Napoleon’s Egyptian

invasion. Napoleon’s troops discovered a stone called Rosetta stone which

contains a hieroglyphic and a demotic version of the same text in parallel with

a Greek translation. A complete decipherment was made by Jean-François

Champollion during 1820 (Allen 1999). Egyptian Hieroglyphs have been

added to the Unicode Standard Version 5.2 in October, 2009.

In 1850 BC, Greeks used a clay disk called Phaistos Disk for secret

message sharing which is shown in Figure 1.2 (a). Size of this disk is about 15

cm in diameter and covered on both sides with a spiral of stamped symbols.

Stamping of symbols was done by pressing pre-formed seals into soft clay, in

a clockwise sequence spiraling towards the disc's center. Subsequently, the

clay was fired at high temperature. The unique feature of Phaistos Disc is that

entire text can be reproduced by inscribing with the help of reusable

characters. There are 241 tokens on the disc, comprising 45 unique signs that

are easily identifiable. It is believed that Phaistos Disc is the first document

of movable type printing (Schwartz 1959). Inspite of several failed attempts

to decipher the Disc, symbols in Phaistos Disc are still a mystery. A set of 46

Phaistos Disc characters, comprising 45 signs and one oblique stroke, have

been encoded in Unicode version 5.1 since April 2008 (Schwartz 1959).

11

(a) Phaistos Disks

(b) Jefferson’s Disk Cipher (c) Enigma Machine

Figure 1.2 Various Ancient Encryption Mechanisms

Substitution cipher was introduced during 600 BC. It is a

cryptographic technique where each letter of plaintext is replaced by a

different letter. Each letter retains its original position in the message, but the

identity is changed. This type of technique was documented during Julius

Caesar's Gallic Wars. Caesar Cipher, Monoalphabetic Ciphers, Hill Cipher,

Playfair Cipher and Polyalphabetic Ciphers are few examples for substitution

12

ciphers. Steganography was introduced by Greek in 600 BC and is used to

write hidden messages. In Steganography, the actual message is hidden using

video, audio, image and text. The word is derived from two Greek words

steganos (covered) and graphein. SIGSALY (also known as the X System,

Project X, Ciphony I, and the Green Hornet) was used in World War II as a

secured speech system for the highest-level allied communications between

then US President Roosevelt and his England counterpart Winston Churchill.

Design of SIGSALY led to the development of digital communications

concepts including transmission of speech using pulse-code modulation.

Two types of rotation ciphers, namely, Jefferson’s disk and Enigma

machine are shown in Figure 1.2(b) and 1.2(c). Jefferson’s disk was

also called as wheel cipher which had 26 wheels. Enigma machine played an

important role during Second World War. It is an encryption and decryption

machine used by Germans to communicate important military messages

secretly.

1.3 GENERAL CLASSIFICATION OF CRYPTOGRAPHY

TECHNIQUES

In general, cryptographic techniques are classified into three

categories, namely, substitution/transposition techniques, key based

techniques, and hash based techniques. Key based techniques are further

classified into symmetric key and asymmetric key encryption algorithms.

Classification of cryptography techniques is shown in Figure 1.3. Some of the

popular algorithms are explained in subsequent subsections.

13

Figure 1.3 Classification of Encryption Algorithms

1.3.1 Substitution Ciphers

The substitution cipher systematically replaces all characters in the

plaintext with other characters. The Caesar cipher was one of the primitive

schemes based on substitution technique. In this method,

each original letter is replaced with some other alphabet which is few

positions apart. One example for Caesar cipher is given below. Consider that

the message, “hello welcome how are you” is to be transmitted. The above

plain text may be transmitted as the following cipher-text “khoor zhofrh krz

duh brx” using the following mapping mechanism.

14

plain : a b c d e f g h i j k l m n o p q r s t u v w x y z

cipher: D E F G H I J K L M N O P Q R S T U V W X Y Z A B C

Later, shift-3 code which is similar to excess-3 code in Boolean

algebra was included in Caesar cipher for communication among generals in

military camps. Table 1.2 presents numerical equivalence of alphabets used in

Caesar cipher.

Table 1.2 Numerical Equivalence of Alphabets Used in Caesar Cipher

a b c d e f g h i j k l m

0 1 2 3 4 5 6 7 8 9 10 11 12

n o p q r s t u v w x y z

13 14 15 16 17 18 19 20 21 22 23 24 25

Each plain text ‘P’ in the message is substituted with a cipher-text

‘C’ using Table 1.2 by employing equation (1.1).

C = E(3, p) = (p + 3) mod 26 (1.1)

The shift may be of any amount ‘K’, which is between 1 to 25,

equation (1.1) can be generalized as

C = E(k, p) = (p + k) mod 26 (1.2)

Decryption is done using equation (1.3)

p = D(k, C) = (Ck) mod 26 (1.3)

Monoalphabetic cipher, Playfair Cipher, Hill Cipher,

Polyalphabetic Cipher and One-Time Pad are some of the popularly used

substitution ciphers.

15

1.3.2 Transposition Ciphers

Transposition cipher rearranges the order of letters in a message by

a constant position shift during encryption. The simplest transposition cipher

is the rail fence technique, in which the plaintext is written as a sequence of

diagonals and then read off as a sequence of rows. For example, the message

"WELCOME ALL" is encrypted as follows using rail fence encryption.

The above message is read as ‘WOLEC MALLE’ through the

normal eye. But the cryptanalyst will be able to decode the above message

correctly as “WELCOME ALL” using rail-fence technique.

1.3.3 Key Based Techniques

History of key based encryption starts from Second World War.

The revolution in computers and electronic communication during the last

century brought lot of opportunities for civilian research into cryptography.

IBM’s former chairman Tomas Watson set up a cryptography research group

during late 90’s in New York. The group was led by Horst Feistel. A private

key encryption system called ‘Lucifer’ was developed by the IBM

cryptography research group. Later, Lucifer was modified and adopted by

National Beauro of standards (NBS) and released as Data Encryption

Standard (DES) during 1977. As explained in previous sections 1.3.1 and

W.. O.. L W.. O.. L

.E .C . M . A . L. .E .C . M . A . L.

..L … E.. ..L … E..

16

1.3.2, for each character in the plaintext, substitution ciphers substitute

another character to generate the cipher-text. Whereas, transposition ciphers

change the position of each character in the plain-text to generate the cipher-

text. However, key based encryptions employ a specific key or password to

generate a cipher-text.

Key based encryption techniques are classified into (i) symmetric

key or secret key encryption and (ii) asymmetric key encryption which are

explained in subsections 1.3.3.1 and 1.3.3.2 respectively. Asymmetric key

encryption employs either a public key or a private key for encryption.

1.3.3.1 Symmetric-key Cryptography

In symmetric-key cryptography both receiver and sender will have

the same key which is used for encryption and decryption. The architecture of

symmetric-key encryption is shown in Figure 1.4. The plaintext is converted

into cipher-text based on a unique key and function.

Figure 1.4 Symmetric Key Encryption

17

The strength of symmetric-key cryptography lies in the non-

reversible function which uses the key to produce cipher text. The following

functions are used for encryption and decryption.

EK(P) = C (1.4)

DK(C) = P

(1.5)

In equation (1.4) and (1.5), EK and DK are encryption and decryption

functions, K is the key, P is the plaintext and C is the ciphertext.

A 64-bit key can give reasonable security but 128 or 256-bit keys

are required for mission critical applications. Increasing the key size improves

security and mitigates security threats such as brute-force, plaintext, chosen

plaintext and differential plaintext attacks. Symmetric-key methods are

further classified into block ciphers and stream ciphers. In block ciphers,

input data is encrypted as block by block, whereas, in stream ciphers, it is

encrypted as bit by bit. In block ciphers, message size will not be altered

whereas, in stream ciphers, it is altered.

1.3.3.2 Asymmetric-key Cryptography

Figure 1.5 shows a typical asymmetric key encryption process. As

the name indicates, two different functions and keys are used by the sender

and the receiver. The keys are: (i) public key which is kept visible and is

transmitted over network and (ii) private key which is kept secret between the

sender and the receiver. Encryption and decryption functions used in

asymmetric cryptography are given in equations (1.6) and (1.7) respectively.

18

Figure 1.5 Asymmetric Key Encryption

EK(P) =C (1.6)

DS(C) = P (1.7)

In equation (1.6), K is the public key which encrypts plaintext P to

generate cipher-text C. In equation (1.7), S is the private key. The private key

‘S’ can be used for encryption and the public key ‘K’ can be used for

decryption and vice versa. Asymmetric key algorithms are slower than

symmetric key algorithms and are prone to security threats. Authentication

using digital signature uses asymmetric key encryption. Table 1.3 lists all

possible key sizes used in symmetric and asymmetric key algorithms.

19

Table 1.3 Key Sizes of Symmetric Key and Asymmetric Key Encryption

Key Size in Symmetric

Key Encryption

Key Size in Asymmetric

Key Encryption

56-bit 384-bit

64-bit 512-bit

80-bit 768-bit

112-bit 1,792-bit

128-bit 2,304-bit

1.3.4 Hash Based Encryption

Cryptographic Hash Function takes a message of any length as

input and produces a fixed length string as output. It is also called as message

digest or a digital fingerprint. A secure hash function is one in which the

original message cannot be deciphered from the given hash (digest) value

other than brute-force technique. Examples of good hash functions are SHA-

512, SHA-256, SHA-160, MD5, RIPEMD and Tiger (Guang 2009), (Rivest

1992). Hashing takes binary or text data as input and creates a constant-length

hash representing a checksum. Different hashing algorithms produce different

sizes of hashes. The original message cannot be recreated using the hash

generated. One-way Unix-based password authentication is based on message

digest. The password is stored as a hash value in a table. During login, the

password is hashed and compared against the stored hash value of the

password. The password is accepted only if both the hashes match. Figure 1.6

shows a model for hash based encryption.

20

Figure 1.6 Hash Based Encryption

1.4 ENCRYPTION ALGORITHMS

Encryption algorithms are used for achieving security and

confidentiality during information sharing. As explained earlier in

Section 1.3, encryption is the process of converting message (plaintext) into

an incomprehensive format called cipher text. The cipher text cannot be easily

understood by an unauthorized user (Min 2002). Two main classifications of

video encryption algorithms are discussed in subsections 1.4.1 and 1.4.2.

1.4.1 Lightweight and Heavyweight Encryption Algorithms

Lightweight algorithms are faster when compared to heavyweight

algorithms as the former employ selective encryption technique (Guido 2003).

Low power devices, networks, designs and operating systems tend to use

lightweight encryption algorithm because of its less computational overhead

and memory utilization. Lightweight algorithms are increasingly used in a

variety of applications, such as wireless and power-constrained devices,

civilian, military and tracking applications, mobile phones, smart cards, and

electronic passports etc. Though, lightweight algorithms are developed for

energy-efficient systems initially, they gain importance because of their

21

widespread usage in pervasive/ubiquitous computing (Ali 2001), (Wang 2008).

On the other hand, heavyweight encryption algorithms provide better security

when compared to lightweight algorithms. Lightweight algorithms attempt to

make a balance between security, performance, and low overall cost in terms of

throughput, area, power consumption, and price in low-resource environments

(Axel 2007). The term lightweight does not necessarily mean weak

cryptographic designs, but it tries to make a better trade-off between the factors

mentioned above. Consequently, they are employed in hand held devices, low

power operating systems and low power wireless sensor networks. Lightweight

algorithms are typically hardware-oriented, and designed to be particularly

compact and efficient. Serpent, Blowfish, Pretty Good Privacy (PGP), Twofish

are few examples for lightweight encryption algorithms.

Heavyweight algorithms are not adopted for real time applications,

mobile devices and wireless networks due to high resource requirements such

as memory, power and computational overhead. Rivest-Shamir-Adelman

(RSA) algorithm and Advanced Encryption Standard (AES) are commonly

employed heavyweight encryption algorithms for video data encryption

(Guido 2003). General steps involved in heavyweight and light weight

encryption algorithms are shown in Figures 1.7 and 1.8.

Figure 1.7 Steps Involved in Heavyweight Encryption

22

Figure 1.8 Steps Involved in Lightweight Encryption

1.4.2 Independent and Joint Encryption Algorithms

Another classification of encryption algorithm is based on how

encryption and compression is done. There are two approaches. In the first

approach, both compression and encryption are done independently as two

different processes by employing suitable and appropriate compression and

encryption algorithms. This approach is called compression-independent

encryption technique and is shown in Figure 1.9. In this approach, only

selective portions of compressed video streams are encrypted and thereby

exploiting the characteristic of the compressed video streams (Fuwen 2005a).

As both compression and encryption are performed separately, this technique

involves high CPU and memory overhead. In the second approach, both

compression and encryption are integrated together as a single process. This

approach is called as joint compression and encryption technique. Two

strategies are followed in joint compression and encryption algorithm. The

first strategy employs encryption after compression, while, the second one

employs encryption before compression. Both strategies are illustrated in

Figure 1.10(a) and 1.10(b). As encryption is done after compression in the

first strategy, we get two-fold advantages, namely, reduction in data size and

execution time. The second strategy encrypts data without compression and

takes more time for encryption. However, both strategies provide two level

security and consume less time when compared to compression independent

encryption algorithms. Secure Motion Picture Experts Group (SECMPEG)

23

(Jurgen 1995), Video Encryption Algorithm (VEA) (Changgui 1998), Real-

time Video Encryption Algorithm (RVEA) (Changgui 1999), are few

examples for joint compression and encryption algorithms. Joint compression

and encryption algorithms are faster in encrypting video data when compared

to independent encryption algorithms.

Figure 1.9 Independent Encryption Technique

(a) Compression before Encryption

(b) Compression after Encryption

Figure 1.10 Joint Compression and Encryption Techniques

1.5 VIDEO ENCRYPTION

Video encryption is the process of converting a plain video into a

cipher video by using a key or hash function which makes it unreadable. The

reverse process of video encryption is called video decryption which converts

the unreadable cipher video into readable plain video using the key or hash

function which has been authenticated during encryption process.

24

1.5.1 Basics of Video Data

Video is a live stream which can be recorded, played or displayed

by information processing devices and real time data processing units, such as

computerized electronic devices (Gordon 2003). The video has been classified

into two categories: (i) linear and (ii) non-linear. The linear video is an active

content stream which has no navigation control. Cinema presentation is one of

the best examples for linear video. The non-linear video stream has user

interactive feature where it supports all controls during navigation. Computer

games, computer based learning and hypermedia are examples of non-linear

video. Video presentations are live streams or recorded file. A recorded video

application can interact with the user through its navigation system, whereas,

a live video stream can interact with the user directly without any interface.

Video is a sequence of frames and its speed can be measured by

using frame rate (Matthias 2008). Number of frame images per unit time is

called frame rate of the video. The frame rate of old mechanical camera

ranges from 6 or 8 frames per second and the frame rate of modern digital

cameras ranges from 120 to more frames per second. The Phase Alternating

Line (PAL) and Sequential Color with Memory (SECAM) standards specify

that frame rate of video camera should be 25 frames/sec and National

Television System Committee (NTSC) specifies 29.97 frame/sec. Film

presentation is one of the interactive video applications which has lower

frame rate of 24 frames/sec. To get the illusion from video frame, the

application should be projected with a minimum frame rate of 15 frames/sec.

25

Figure 1.11 Aspect Ratio of Traditional Television in RGB

Depending on the device, different dimensions are used for color

and gray scale video signals. Aspect Ratio (AR) is used to measure

dimensions of the video screen. It is the ratio of screen width and height.

Figure 1.11 shows the Aspect Ratio (AR) of traditional television. The aspect

ratio of a traditional television screen is 1:33:1 and High Definition (HD)

television is 1:78:1.

The color representation of a video is called its color model. There

are various types of color models used in video display system. In NTSC and

PAL standards YIQ color model which is similar to YUV color model is used

whereas in SECAM standard, YDbDr color model is used. Color

representation of UV color model is shown in Figure 1.12.

Figure 1.12 U-V Color Plane

26

1.5.2 RGB and YUV Representation of Video Signals

In general, the Red, Green, Blue (RGB) color system is used in

representation of color images and video (Iain 2003). Appropriate

contribution of RGB can generate any desired color. In general, a color is

represented using luminance and chrominance. Let Y be the luminance which

represents the brightness of the color. Luminance can be calculated by using

the weighted sum of three colors R, G and B. Color difference of these color

models Cr, Cg and Cb can be calculated by subtracting the luminance from

each primary component using the following expressions:

Cr = Wr* (R - Y)

(1.8)

Cg=Wg* (G - Y) (1.9)

Cb = Wb* (B - Y) (1.10)

where, Wr, Wg and Wb are weights of R, G and B.

Among these three color differences Cr, and Cg are linearly

independent. The color difference Cb can be expressed as a linear

combination of the two linearly independent signals R and G. Luminance Y

and any of the two color differences are used to represent a color. Some of the

popular standards such as NTSC, PAL and SECAM use three components,

luminance Y, blue color difference U and red color difference V to represent a

color (Iain 2003). This is called YUV color system. The RGB color signal can

be converted into YUV model by using the following expressions:

Y = (0.257 * R) + (0.504 * G) + (0.098 * B) +16 (1.11)

27

U = (0.439 * R) + (0.368 * G) – (0.071 * B) +128 (1.12)

V = -(0.148 * R) - (0.291 * G) + (0.439 * B) +128 (1.13)

Human Visual System (HVS) is less sensitive to chrominance than

brightness. This is because chrominance signals are represented by lower

resolution than luminance. Therefore, YUV is preferred over RGB in some

standards. Video data processing quality is very important parameter and is

used to measure signal strength of video signal. Subjective video quality (Ke

1997), (Kotevski 2010), (Maria 2012) (Andrew 1999) for video processing

system is evaluated as follows:

a) Select all video sequences for testing

b) Define a system for evaluation

c) Define a method for presenting video sequences to experts

d) Invite enough number of experts

e) Do testing on video

f) Collect expert’s ratings

g) Calculate average points for each testing based on experts

rating

1.5.3 Challenges in Video Encryption

Internet users have been experiencing viruses, trojans, hacking and

data espionage. The greatest risk for internet users lies in infection of

computers with malware and spyware. The economic losses incurred due to

officially recorded Internet crimes in Germany amounted to approximately

61.5 million euro dollars in 2010.

28

With the increasing trend in digital communication and multimedia

commerce applications, it is important to provide security to audio, video and

other information involved in transmission. In addition, the need for video has

been growing rapidly in areas such as education, communication, publishing

and entertainment in recent years. Combining digital video, database and

communication network technologies enhance the ability of delivering video

along with text, images and audio through networks. Applications such as

digital libraries, video databases, video-telephony/conferencing, and Video-

on-Demand (VoD) are already in use. However, improvement is required in

terms of storing huge amount of video data at relatively low cost, efficient

organization of video data, retrieving, delivering, and presenting the requested

data for easy access for the above mentioned services. One of the major

challenges in deploying digital video is the huge volume of uncompressed

video which may overwhelm the available communication channels and

storage systems.

Two important factors for video data encryption are file size and

speed of execution. A video of three hours movie takes nearly 1 GB size. Real

time applications such as video conferencing and video telephony require

more memory while in operation. It consumes more time to process such

large amount of data for storing or transferring through a network. This is due

to the large size of video files when compared to text files. Therefore, video

data need to be compressed before transmitting or sharing through the

network.

Video is the primary commodity in the world of E-Commerce. As

technology advances and access to markets expand, the need to protect video

29

data, to ensure confidentiality, integrity, and availability for making critical

personal, business, or government decisions became more important. Without

protection of video data, success of an E-Commerce marketplace may be

short-lived. Video security plays a major role in providing threat free

communication. Implementing weak video security mechanisms can result in

the loss of trust, reputation, and money for consumers, businesses, and

governments. The sharp increase in the number of fraud, extortion, and

identity theft crimes is the primary result of weak video security mechanisms.

The cost of implementing basic mechanisms to protect video data is generally

much less expensive than a security breach. Weak video security mechanisms

can generally be attributed to the lack of understanding and acknowledgment

of potential threats to information.

1.6 VIDEO COMPRESSION

With the growth of social and professional networks such as

Facebook, YouTube and LinkedIn etc, large volumes of video files are shared

and transferred across the globe. However, video downloads experience

significantly slow transfer rate. Use of compression techniques before

transferring or sharing large video files will reduce transfer time and further

reduce the cost of communication (Guoping 2004).

Compression is done to minimize the number of bits required to

represent data in memory. The simplest form of data compression is sampling

of band-limited images. In this method, an infinite number of pixels per unit

area is reduced to one sample without any loss of information. Data

30

compression is applied primarily during storage of information and

transmission. Video transmission methods are used in applications such as

remote sensing via satellite, broadcast television, military communications via

aircraft, sonar, radar and teleconferencing. Data compression aids in

development of faster algorithms as number of operations is significantly

reduced for compressed data (Anil 1989).

Usually, television video still image has spatial resolution of

approximately 512 x 512 pixels per frame. Assuming 8-bit per pixel for color

channel and 30 frames in a second, this will translate into a rate of 180 x 106

bits/sec. Depending on the application, digital video still image data rate vary

from 105-bit per frame to 108-bit per frame. Large-channel capacity and

memory requirements for digital video storage and transmission have made it

necessary to employ compression methods.

Converting an analog video signal into a digital signal will result in

increased bandwidth requirement for transmission. Consider a 4 MHz

television signal sampled at Nyquist rate with 8-bit per sample. It would

require a bandwidth of 32 MHz when transmitted using a digital modulation

scheme such as Phase Shift Keying (PSK). It requires 1 Hz for every 2-bit.

Digitized representation has more advantages over its analog counter part in

terms of processing flexibility, random access in storage and higher signal to

noise ratio for transmission with the possibility of errorless communication

etc. These advantages come with a price tag of eightfold increase in

bandwidth. Data compression techniques are employed to reduce the

bandwidth of digital video signal (Anil 1989).

31

Table 1.4 History of International Video Compression Standards

Year Standard Publisher Popular Implementations

1988 H.261 ITU-TVideoconferencing, Video Telephony

1993 MPEG-1 Part 2 ISO, IEC Video-CD

1995H.262/MPEG-2

Part 2ISO, IEC, ITU-

TDVD, Digital Video Broadcasting, SVCD

1996 H.263 ITU-TVideoconferencing, Video Telephony, Video on Mobile Phones (3G Phone)

1999 MPEG-4 Part 2 ISO, IEC Video on Demand

2002H.264/MPEG-4

AVCISO, IEC, ITU-

T

Digital Video Broadcasting, iPod Video, Apple TV, HD DVD

2008 VC-2 (Dirac) ISO, BBCVideo on Demand, HDTV broadcast, UHDTV

Data compression methods fall into two categories: (i) predictive

coding and (ii) transform coding. Predictive coding exploits the redundancy in

the data. As the name indicates, transform coding, transforms the input video

frame into different arrays and packs them into a number of small samples.

Other compression algorithms are either generalizations or combinations of

predictive coding and transform coding. It is likely that video compression

results in some distortion of data because of analog to digital conversion and

omission of some insignificant information. However, efficient video

compression techniques try to reduce distortion. The timeline of video

compression techniques and standards are tabulated in Table 1.4.

32

1.6.1 Lossy and Lossless Video Compression

Basically, compression packs the input information into a smaller

space. Without compression, video data require huge memory space. Hence

video must be compressed before it is encrypted, transmitted, stored or put up

on the web. Compression techniques are classified into (i) lossy compression

and (ii) lossless compression based on the quality of the output generated

(Alexandre 2003). In lossy compression, files are reduced to 15% - 30% of

their original size. As significant portion of data is lost, quality of output after

decompression may not be the same in lossy compression. MPEG-2 is a lossy

compression technique used in Digital Versatile Disk’s (DVDs). It reduces

files at least fifteen times smaller, but the end user perceives DVDs as having

high-quality video (Fernando 2011), (Nam 2010). In Lossless compression,

there is no significant difference in file sizes before and after compression. It

is mainly used for text and data files where high accuracy is required such as

bank records, employee details and text articles.

1.7 SUMMARY OF RESEARCH CONTRIBUTIONS

With the advent of ubiquitous computing, information sharing

among heterogeneous devices and groups demands high security with low

resource requirements. This in turn necessitates highly efficient and effective

encryption algorithms to share information without the fear of being

eavesdropped. Lightweight encryption algorithms are preferred among low

power devices, whereas, heavyweight algorithms are preferred when security

is the top priority irrespective of high resource requirements. This thesis

proposes three efficient video encryption algorithms which are briefly

described in the following subsections.

33

1.7.1 Fast Random Bit Encryption (FRBE) Technique

Fast Random Bit Encryption (FRBE) compresses the input video

data using GZIP and then encrypts the compressed video by employing

various strategies that give multilevel security to video data that is being

shared. FRBE is a lossless, lightweight, joint compression and encryption

technique. The computing overhead and memory utilization using the

proposed method, FRBE, is less than that of the existing algorithms such as

Advanced Encryption Standard (AES), RSA, Pretty Good Privacy (PGP)

and Tiny Encryption Algorithm (TEA) (Guido 2003), (Yongcheng 1996),

(David 1994).

1.7.2 Puzzle Fast Random Bit Encryption (PFRBE) Technique

Puzzle Fast Random Bit Encryption (PFRBE) is the enhanced

version of FRBE. It uses Entropy Coding (EC), Puzzle Transform (PT) and

Randomized Arithmetic Coding (RAC) for compression. Then, Fast Random

Bit Encryption (FRBE) is employed for encryption. This is a lossless, joint

compression and encryption technique. PFRBE takes less time for encryption,

because the input video is compressed at multiple levels. It is observed that

PFRBE gives better results in terms of compression efficiency and memory

utilization when compared to existing algorithms such as SECMPEG, VEA,

RVEA, ZigZag (Jurgen 1995), (Changgui 1998a), (Changgui 1999), (Lei

1996).

1.7.3 Enhanced Randomized Arithmetic Coding (ERAC)

Enhanced Randomized Arithmetic Coding (ERAC) incorporates

interval splitting Randomized Arithmetic Coding (RAC) (Grangetto 2004)

with a simple bit–wise XOR operation and randomly generated keys for

34

encryption. In this work, randomization is coupled with an XOR operation to

further scramble the video in order to achieve high degree of security. This

technique does not compromise on coding efficiency and mitigates most of

the cryptanalytic attacks such as cipher-only, chosen-plaintext and the chosen-

cipher text attacks. This thesis also presents the comparative analysis based on

compression and encryption efficiency by employing the above mentioned

algorithms.

1.8 ORGANIZATION OF THE THESIS

Rest of the thesis is organized as follows: the literature review is

presented in Chapter 2. Details of Fast Random Bit Encryption (FRBE), one

of the proposed lightweight encryption techniques are discussed in Chapter 3.

Chapter 4 presents the methodology used in implementing another video

encryption algorithm called Puzzle Fast Random Bit Encryption (PFRBE).

PFRBE is the enhanced version of FRBE. An improved and efficient

lightweight encryption technique based on arithmetic coding, called Enhanced

Randomized Arithmetic Coding (ERAC), is explained in Chapter 5.

Comparative analysis of the above mentioned techniques are summarized in

Chapter 6. Chapter 7 concludes the thesis with the directions for future work.