Lossless Information Hiding in Images

1st Edition Zhe-Ming Lu

Visit to download the full and correct content document:
https://ebookmass.com/product/lossless-information-hiding-in-images-1st-edition-zhe-
ming-lu/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Seeking Truth and Hiding Facts: Information, Ideology,

and Authoritarianism in China Jeremy L. Wallace

https://ebookmass.com/product/seeking-truth-and-hiding-facts-
information-ideology-and-authoritarianism-in-china-jeremy-l-
wallace/

Chinese National Identity in the Age of Globalisation

1st ed. Edition Lu Zhouxiang

https://ebookmass.com/product/chinese-national-identity-in-the-
age-of-globalisation-1st-ed-edition-lu-zhouxiang/

Eco-Cities and Green Transport 1st Edition Huapu Lu

https://ebookmass.com/product/eco-cities-and-green-transport-1st-
edition-huapu-lu/

Search Engines: Information Retrieval in Practice 1st

Edition Croft

https://ebookmass.com/product/search-engines-information-
retrieval-in-practice-1st-edition-croft/
Topological Insulator and Related Topics 1st Edition Lu
Li (Editor)

https://ebookmass.com/product/topological-insulator-and-related-
topics-1st-edition-lu-li-editor/

Hiding in Plain Sight: Deceptive Tactics and the

Criminal Victimization Process 1st ed. 2020 Edition
James F. Kenny

https://ebookmass.com/product/hiding-in-plain-sight-deceptive-
tactics-and-the-criminal-victimization-process-1st-
ed-2020-edition-james-f-kenny/

A Woman in Hiding (A Woman Lost Book 10) T B Markinson

https://ebookmass.com/product/a-woman-in-hiding-a-woman-lost-
book-10-t-b-markinson/

COVID-19: The Essentials of Prevention and Treatment

1st Edition Jie-Ming Qu

https://ebookmass.com/product/covid-19-the-essentials-of-
prevention-and-treatment-1st-edition-jie-ming-qu/

Images of the Past 7th Edition, (Ebook PDF)

https://ebookmass.com/product/images-of-the-past-7th-edition-
ebook-pdf/
Lossless Information
Hiding in Images
This page intentionally left blank
Lossless Information
Hiding in Images

Zhe-Ming Lu
Professor, School of Aeronautics and Astronautics
Zhejiang University
Hangzhou, P.R. China

Shi-Ze Guo
Professor, School of Computer Science
Beijing University of Posts and Communications
Beijing, China

AMSTERDAM • BOSTON • HEIDELBERG • LONDON

NEW YORK • OXFORD • PARIS • SAN DIEGO
SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO

Syngress is an imprint of Elsevier

Syngress is an imprint of Elsevier
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
Copyright © 2017 Zhejiang University Press Co., Ltd., published by Elsevier Inc.
All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage
and retrieval system, without permission in writing from the publisher. Details on how to
seek permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the
Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by
the Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional
practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described
herein. In using such information or methods they should be mindful of their own safety
and the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or
editors, assume any liability for any injury and/or damage to persons or property as a
matter of products liability, negligence or otherwise, or from any use or operation of any
methods, products, instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-12-812006-4

For information on all Syngress publications

visit our website at https://www.elsevier.com/

Publisher: Todd Green

Acquisition Editor: Simon Tian
Editorial Project Manager: Naomi Robertson
Production Project Manager: Priya Kumaraguruparan
Cover Designer: Mark Rogers

Typeset by TNQ Books and Journals

Contents

Preface .................................................................................................. xi

CHAPTER 1 Introduction ............................................................ 1

1.1 Background...................................................................... 1
1.1.1 Definition of Images .................................................. 1
1.1.2 Image Processing and Image Analysis .......................... 4
1.1.3 Network Information Security ..................................... 8
1.1.4 Image Protection and Image Authentication................. 10
1.2 Overview of Information Hiding .........................................12
1.2.1 Basic Concepts Related to Information Hiding ............. 13
1.2.2 Properties and Requirements of Information Hiding ...... 15
1.2.3 Information Hiding Models....................................... 17
1.2.4 Research Branches of Information Hiding ................... 19
1.2.5 Classification of Information Hiding Technologies ........ 22
1.2.6 Applications of Information Hiding Techniques............ 24
1.3 Overview of Image Coding and Compression Techniques........27
1.3.1 Source Coding and Data Compression ........................ 27
1.3.2 Lossless Image Coding Techniques ............................ 27
1.3.3 Lossy Image Coding Techniques................................ 29
1.3.4 Vector Quantization ................................................. 30
1.3.5 Block Truncation Coding.......................................... 31
1.3.6 JPEG..................................................................... 32
1.3.7 JPEG2000.............................................................. 35
1.4 Overview of Information Hiding Techniques for Images .........38
1.4.1 Robust Image Watermarking ..................................... 38
1.4.2 Fragile Image Watermarking ..................................... 43
1.4.3 Image Fingerprinting ............................................... 45
1.4.4 Image Steganography............................................... 47
1.4.5 Lossless Information Hiding in Images ....................... 50
1.5 Applications of Lossless Information Hiding in Images...........57
1.5.1 Lossless Hiding Authentication.................................. 57
1.5.2 Fidelity Control for Cover Media ............................... 59
1.5.3 Status Recovery for Cover Media............................... 60
1.6 Main Content of This Book................................................62
References.............................................................................63

v
vi Contents

CHAPTER 2 Lossless Information Hiding in Images on

the Spatial Domain.................................................69
2.1 Overview of Spatial DomaineBased Information Hiding ........69
2.2 Modulo AdditioneBased Scheme .......................................71
2.2.1 Embedding Process.................................................. 71
2.2.2 Authentication Process ............................................. 71
2.2.3 Explanation of the Modulo Addition Operation ............ 71
2.3 Difference ExpansioneBased Schemes ................................73
2.3.1 Tian’s Scheme ........................................................ 73
2.3.2 Alatter’s Scheme ..................................................... 76
2.4 Histogram ModificationeBased Schemes .............................80
2.4.1 Original Histogram ShiftingeBased Scheme................ 80
2.4.2 Adjacent Pixel DifferenceeBased Scheme................... 80
2.4.3 Multilevel Histogram ModificationeBased Scheme....... 82
2.4.4 Hybrid Prediction and Interleaving Histogram
ModificationeBased Scheme..................................... 93
2.5 Lossless CompressioneBased Schemes.............................. 102
2.5.1 Lossless Bit-Plane Compression in the Spatial
Domain ................................................................102
2.5.2 Lossless RS-Data Embedding Method........................104
2.5.3 Lossless G-LSB Data Embedding Method ..................106
2.5.4 Look-up Table-Based Scheme for Error Diffused
Halftone Images.....................................................110
2.6 Reversible Secret SharingeBased Schemes......................... 119
2.6.1 Data Hiding in Reversible Secret Sharing ...................120
2.6.2 Joint Secret Sharing and Data Hiding for
Block Truncation Coding-Compressed Images ............129
2.7 Summary ...................................................................... 137
References........................................................................... 139
CHAPTER 3 Lossless Information Hiding in Images
on Transform Domains .......................................... 143
3.1 Introduction................................................................... 143
3.1.1 Invertible Mapping .................................................143
3.1.2 Requirements of Reversible Information Hiding
to Integer Transforms..............................................145
3.2 Overview of Transform-Based Information Hiding............... 146
3.2.1 Discrete Cosine TransformeBased Information
Hiding..................................................................147
 Contents vii

3.2.2 Discrete Fourier TransformeBased Information

Hiding..................................................................148
3.2.3 Discrete Wavelet TransformeBased Information
Hiding..................................................................149
3.3 Integer Discrete Cosine TransformeBased Schemes............. 151
3.3.1 Integer Discrete Cosine Transform ............................151
3.3.2 Integer Discrete Cosine TransformeBased Lossless
Information Hiding Scheme Using Companding
Technique .............................................................157
3.3.3 Histogram Shift Technique in Integer Discrete
Cosine Transform Domain .......................................172
3.3.4 Lossless Information Hiding by Adaptive Coefficient
Modification in Integer Discrete Cosine Transform
Domain ................................................................179
3.4 Integer Wavelet TransformeBased Schemes ....................... 189
3.4.1 Value Expansion Technique in the Integer Wavelet
Transform Domain .................................................190
3.4.2 Coefficients’ Magnitude Prediction Technique
in Integer Wavelet Transform Domain .......................193
3.4.3 Comparisons Between Generalized Lossless
CompressioneBased and Value ExpansioneBased
Schemes ...............................................................195
3.5 Summary ...................................................................... 201
References........................................................................... 202
CHAPTER 4 Lossless Information Hiding in Vector
Quantization Compressed Images ......................... 205
4.1 Introduction................................................................... 205
4.1.1 Basic Framework of Lossless Information Hiding
in the Compressed Domain ......................................205
4.1.2 Requirements of Lossless Information Hiding
in the Compressed Domain ......................................205
4.1.3 Vector Quantization ................................................207
4.2 Overview of Vector QuantizationeBased Information
Hiding.......................................................................... 208
4.2.1 Overview of Vector QuantizationeBased Image
Watermarking ........................................................209
4.2.2 Overview of Vector QuantizationeBased Lossless
Information Hiding.................................................211
4.3 Modified Fast Correlation Vector QuantizationeBased
Scheme......................................................................... 212
viii Contents

4.3.1 Lossless Information Hiding for Basic Vector

Quantization..........................................................213
4.3.2 Fast Correlation Vector QuantizationeBased Image
Compression..........................................................214
4.3.3 Modified Fast Correlation Vector QuantizationeBased
Lossless Information Hiding.....................................215
4.4 Side Match Vector QuantizationeBased Schemes ................ 219
4.4.1 Side Match Vector Quantization................................219
4.4.2 Chang and Wu’s Scheme .........................................221
4.4.3 Chang and Lu’s Scheme ..........................................222
4.4.4 ChangeTaieLin’s Scheme.......................................232
4.5 Vector QuantizationeIndex CodingeBased Schemes............ 238
4.5.1 Joint Neighboring CodingeBased Scheme..................238
4.5.2 Vector QuantizationeIndex Residual Value
CodingeBased Scheme ...........................................242
4.5.3 Path Optional Lossless Information Hiding Scheme .....247
4.5.4 Difference CodingeBased Scheme............................257
4.5.5 Improved Joint Neighboring CodingeBased Scheme....261
4.5.6 Lossless Hiding Scheme for Two-Stage Vector
Quantization Compressed Images..............................269
4.6 Summary ...................................................................... 276
References........................................................................... 276
CHAPTER 5 Lossless Information Hiding in Block Truncation
CodingeCompressed Images ................................ 279
5.1 Block Truncation Coding................................................. 279
5.1.1 Original Block Truncation Coding Algorithm..............279
5.1.2 Absolute Moment Block Truncation Coding ...............281
5.2 Overview of Block Truncation CodingeBased Information
Hiding.......................................................................... 282
5.2.1 Overview of Block Truncation CodingeBased Image
Watermarking ........................................................282
5.2.2 Overview of Block Truncation CodingeBased
Lossless Information Hiding.....................................290
5.3 Bitplane FlippingeBased Lossless Hiding Schemes ............. 290
5.3.1 Original Bitplane FlippingeBased Scheme .................291
5.3.2 Improved Bitplane FlippingeBased Scheme ...............292
5.3.3 Bitplane FlippingeBased Scheme With Histogram
Shifting of Mean Tables ..........................................292
5.4 Mean CodingeBased Lossless Hiding Schemes................... 297
5.4.1 Prediction-Error ExpansioneBased Scheme................297
 Contents ix

5.4.2 Joint Neighboring CodingeBased Scheme..................301

5.4.3 Blockwise CodingeBased Scheme............................310
5.5 Lossless Data Hiding in Block Truncation
CodingeCompressed Color Images ................................... 319
5.5.1 Block Truncation Coding Compression for
Color Images.........................................................320
5.5.2 Difference ExpansioneBased Scheme........................321
5.5.3 Blockwise CodingeBased Scheme............................324
5.6 Summary ...................................................................... 329
References........................................................................... 330
CHAPTER 6 Lossless Information Hiding in JPEG-
and JPEG2000-Compressed Images ...................... 333
6.1 Introduction................................................................... 333
6.1.1 JPEG....................................................................333
6.1.2 JPEG2000.............................................................336
6.1.3 Challenges ............................................................338
6.2 Lossless Information Hiding in JPEG Images...................... 339
6.2.1 Overview of Information Hiding in JPEG Images ........339
6.2.2 Fridrich et al.’s Invertible Watermarking Methods
for Image Authentication.........................................340
6.2.3 Fridrich et al.’s Lossless Embedding Method
With File Size Preservation......................................343
6.2.4 Xuan et al.’s Histogram PaireBased Lossless
Data Embedding Scheme.........................................346
6.2.5 Qian and Zhang’s Scheme .......................................352
6.2.6 Joint Modification of Quantization Table and Discrete
Cosine Transform CoefficientseBased Scheme ...........353
6.2.7 Variable Length Code MappingeBased Scheme ..........370
6.3 Lossless Information Hiding in JPEG2000 Images ............... 384
6.3.1 Overview ..............................................................384
6.3.2 Chen et al.’s Irreversible Watermarking Scheme ..........386
6.3.3 Bit-Depth EmbeddingeBased Scheme .......................390
6.3.4 Reversible Information Hiding Based on Histogram
Shifting for JPEG2000 Images..................................394
6.4 Summary ...................................................................... 399
References........................................................................... 400

Index...................................................................................................403
This page intentionally left blank
Preface

The enormous popularity of the World Wide Web in the early 1990s demonstrated the
commercial potential of offering multimedia resources through digital networks.
Representation of media in digital format facilitates its access. Digital media
includes text, digital audio, images, video, and software. The recent growth of net-
worked multimedia systems has increased the need for the protection of digital
media. Since commercial interests seek to use the digital networks to offer digital
media for profit, it is particularly important for the protection and enforcement of
intellectual property rights, and they have a strong interest in protecting their owner-
ship rights. On the other hand, the age of digital multimedia has brought many advan-
tages in the creation and distribution of information. Representation of media in
digital format enhances the accuracy, efficiency, and portability of existence of
data. The powerful multimedia manipulation software has made it possible to edit
and alter the media’s content seamlessly. Since the ease of copying and editing
decreases the credibility of multimedia, a secure authentication system is needed to
verify data integrity and trustworthiness. Furthermore, the rapid development of
the Internet requires confidential information that needs to be protected from the un-
authorized users. Thus, the standard and concept of “what you see is what you get
(WYSIWYG),” which we encounter sometimes while printing images or other mate-
rials, is no longer precise and would not fool a steganographer as it does not always
hold true. Images can be more than what we see with our human visual system (HVS);
hence, they may convey more than merely 1000 words. For decades, people strove to
develop innovative methods for secret communication.
Under these circumstances, many approaches have been presented to protect the
digital media itself or utilize the digital media to protect other important informa-
tion. These approaches can be mainly classified into two categories, i.e., cryptog-
raphy and information hiding. In conventional cryptographic systems, the sender
generates a digital signature for an image in advance using a public key cryptog-
raphy system such as the RivesteShamireAdleman system. The sender then trans-
mits both the digital signature and the corresponding image to the receiver. Later, the
receiver can verify the integrity and authenticity of the received image by using the
corresponding digital signature. The cryptographic system permits only valid key-
holders access to encrypted data, but once such data is decrypted there is no way
to track its reproduction or retransmission. Information hiding, which is also known
as data hiding, is distinct from cryptography as it aims to make the embedded data
unrecoverable and inviolateable. Information hiding is a method of hiding secret
messages into a cover medium so that an unintended observer will not be aware
of the existence of the hidden messages.
Information hiding techniques can be classified into three techniques, i.e., steg-
anography, watermarking, and fingerprinting. These techniques are quite difficult to
tease apart especially for people coming from different disciplines. The term steg-
anography is retrieved from the Greek words stegos means cover and grafia meaning
xi
xii Preface

writing, defining it as covered writing. The similarity between steganography and

cryptography is that both are used to conceal information. However, the difference
is that steganography does not reveal any suspicion about the hidden information to
the user. Therefore the attackers will not try to decrypt information. There are other
two techniques that seem to be same as steganography. They are watermarking and
fingerprinting. Both these techniques sounds to be the same and provide the same
end goals, but both are very different in the way they work. Watermarking allows
a person to provide hidden copyright notices or other verification licenses, whereas
fingerprinting uses each copy of the content and makes it unique to the receiver.
Watermarking is usually a signature to identify the origin, and all the copies are
marked in the same way. However, in fingerprinting different unique copies are
embedded for distinct copies. Digital watermarking has also been proposed as a
possible solution for data authentication and tamper detection. The invisible authen-
ticator, sensitive watermark is inserted using the visual redundancy of HVS and is
altered or destroyed when the host image is modified by various linear or nonlinear
transformations. The changes of authentication watermark can be used to determine
the modification of the marked image, and even locate the tampered area. Because
the watermark is embedded in the content of image, it can exert its efficiency in the
whole lifecycle.
Today, there is a huge volume of literature on information hiding techniques.
However, most of the existing information hiding schemes distort the original image
irreversibly; then, the challenge becomes one of minimizing distortion relative to ca-
pacity. In several applications, such as medical or military imaging, any distortion is
intolerable. In such cases, lossless information hiding schemes are the only recourse.
To meet this need, the concept of distortion-free embedding has become a very
important issue, especially in sensitive images. Lossless information hiding, also
called reversible information hiding or reversible data hiding, allows a sender to
embed invisible information into an original image in a reversible manner. Then,
the receiver can extract the embedded data and restore the original image. Lossless
information hiding in images is gaining more attention in the past few years because
of its increasing applications in military communication, health care, and law
enforcement.
Lossless information hiding can be used in many applications. A possible appli-
cation is to use lossless reversible watermarking algorithms to achieve the lossless
watermark authentication, supporting completely accurate authentication for the
cover media, which is actually the original intention of reversible watermarking
schemes. In some applications, people are more concerned about the quality of
the cover media itself. In this type of application, the common requirements are
that the watermarking algorithm and the watermark embedding process do not intro-
duce permanent distortion to the cover media. A special class of application that we
are most likely to think of is the special accuracy requirement for special media, such
as medical images, military images, remote sensing images, legal evidence images,
secret documents, precious works of art, and science experimental images. For this
type of application, even 1-bit permanent loss in the original cover media is not
 Preface xiii

allowed, so the data embedding algorithms must be reversible. Since reversible

watermark embedding algorithms can remove the embedding distortion completely,
they can be referred to as data embedding styles with 100% fidelity. Another appli-
cation is to restore the image modification operations. In some image processing ap-
plications, the process is completed by a few simple adjustments. If the image
processing operator worries about the fact that the users are not satisfied with the
image processing results, he can treat the parameters as a watermark and reversibly
embed it in the cover image and in the future restore the image to its original state or
its approximate state, thus you do not have to store a lot of original images.
Lossless information hiding algorithms for images can be classified into three
categories, i.e., the spatial domainebased, the transform domainebased, and the
compressed domainebased schemes. Transform domainebased lossless informa-
tion hiding methods can be classified into two categories, i.e., integer discrete cosine
transform (DCT)ebased schemes and integer discrete wavelet transform (DWT)e
based schemes. Here, the transform used in reversible information hiding should
be in the category of invertible mapping. Nowadays, since more and more images
are being stored in compressed formats such as JPEG and JPEG2000 or transmitted
based on vector quantization (VQ) and block truncation coding (BTC), more and
more efforts are being focused on developing the compressed domain information
hiding approaches. Here, we view the compressed image as the cover image, and
reversibility means that after extracting the secret data from the unattacked stego im-
age, we can recover the original compressed image losslessly.
In general, payload capacity, stego image quality, and complexity of the data
embedding algorithm are the three major criteria used to evaluate lossless informa-
tion hiding. Payload capacity means how much secret information can be embedded
in an image. The quality of a stego image is measured by the peak signal-to-signal
ratio (PSNR): a higher PSNR value can guarantee less distortion caused in the cover
image. Moreover, the complexity of the data embedding algorithm should be as sim-
ple as it is effective. In practice, high payload capacity and low distortion are con-
flicting requirements: the larger the capacity created by a reversible embedding
technique, the higher is the distortion introduced by embedding.
Based on this background, this book is devoted to lossless information hiding
techniques for images. This book is suitable for researchers in the field of informa-
tion hiding. Our aim is to recommend a relatively novel monograph to cover recent
progress of research hotspots in the area of lossless information hiding such that the
researchers can refer to this book to have a comprehensive understanding and carry
out in-depth research in their own directions. This book consists of six chapters.
Chapter 2 discusses the lossless information techniques in the spatial domain. In
this chapter, we first overview the spatial domainebased lossless schemes. Then,
we discuss some typical spatial domain lossless information hiding methods.
Chapter 3 discusses transform domainebased lossless schemes. In this chapter,
we first introduce some related concepts and requirements for lossless information
hiding in transform domains. Then we give a brief overview of transform domaine
based information hiding. Next we introduce two types of transform domainebased
xiv Preface

lossless information hiding methods. Chapter 4 focuses on the VQ-based lossless in-
formation hiding schemes. We first review the schemes related to VQ-based infor-
mation hiding. Then, we mainly focus on three kinds of VQ-based lossless
information hiding schemes. Chapter 5 discusses the topic of embedding data in
BTC-compressed images with lossless hiding techniques. First, we introduce the
block truncation coding technique. Then, we review the schemes related to BTC-
based information hiding. Then, we focus on two topics, i.e., lossless information
hiding schemes for BTC-compressed grayscale images and color images. Chapter
6 first introduces JPEG and JPEG2000 compression techniques in brief, together
with the embedding challenges. Then, we introduce the lossless information hiding
schemes for JEPG- and JPEG2000-compressed images.
This book is a monograph in the area of information hiding. It focuses on one
branch of the field of information hiding, i.e., lossless information hiding. Further-
more, it focuses on the most popular media, images. This book embodies the
following characteristics. (1) Novelty: This book introduce many state-of-the-art
lossless hiding schemes, most of that come from the authors’ publications in the
past 5 years. The content of this book covers the research hotspots and their recent
progress in the field of lossless information hiding. After reading this book, readers
can immediately grasp the status, the typical algorithms, and the trend in the field of
lossless information hiding. For example, in Chapter 6, reversible data hiding in
JPEG2000 images is a very new research branch. (2) All roundedness: In this
book, lossless information hiding schemes for images are classified into three cate-
gories, i.e., spatial domainebased, transform domainebased, and compressede
domain based schemes. Furthermore, the compressed domainebased methods are
classified into VQ-based, BTC-based, and JPEG/JPEG2000-based methods. Espe-
cially, the lossless information hiding in JPEG images is very useful since most of
the images are stored in the JPEG format. Therefore, the classification of lossless
hiding schemes covers all kinds of methods. (3) Theoretical: This book embodies
many theories related to lossless information hiding, such as image compression,
integer transforms, multiresolution analysis, VQ, BTC, JPEG, and JPEG2000. For
example, in Chapter 3, several definitions related to invertible mappings and integer
DCT transforms are introduced in detail to understand the content of later chapters
easily. (4) Practical: It is suitable for all researchers, students, and teachers in the
fields of information security, image processing, information hiding, and communi-
cations. It can guide the engineers to design a suitable hiding scheme for their spe-
cial purpose, such as copyright protection, content authentication, and secret
communication in the fields of military, medicine, and law.
This book is completely written by Prof. Zhe-Ming Lu. The research fruits of this
book are based on the work accumulation of the author for over a decade, most of
which comes from the fruits of PhD and master students supervised by Prof. Lu.
For example, Dr. Zhen-Fei Zhao and Dr. Hao Luo carried out the research work
on reversible secret sharingebased lossless information hiding schemes supervised
by Prof. Lu. Dr. Bian Yang, who was a former masters and PhD student, cosuper-
vised by Prof. Lu, carried out the research work in Germany on lossless information
 Preface xv

hiding schemes based on integer DCT/DWT transforms as the main part of his thesis
topic. Dr. Yu-Xin Su, who was a former masters student, supervised by Prof. Lu, car-
ried out the research work on lossless information hiding schemes for BTC-
compressed color images as part of his thesis topic. Mr. Xiang Li, who was a former
masters student, supervised by Prof. Lu, carried out the research work on lossless
information hiding in JPEG/JPEG2000-compressed images as part of his thesis
topic. We would like to show our great appreciation of the assistance from other
teachers and students at the Institute of Astronautics Electronics Engineering of
Zhejiang University. Part of research work in this book was supported by the
National Scientific Foundation of China under the grants 61171150 and 61003255
and the Zhejiang Provincial Natural Science Foundation of China under the grants
R1110006 and RLY14F020024. Owing to our limited knowledge, it is inevitable
that errors and defects will appear in this book, and we adjure readers to criticize.

Zhe-Ming Lu
Hangzhou, China
June 2016
This page intentionally left blank
 CHAPTER

Introduction
1
1.1 BACKGROUND
1.1.1 DEFINITION OF IMAGES
1.1.1.1 Images
An image is a visual representation of something that depicts or records visual
perception. For example, a picture is similar in appearance to some subject, which
provides a depiction of a physical object or a person. Images may be captured by
either optical devices, such as cameras, mirrors, lenses, and telescopes, or natural
objects and phenomena, such as human eyes or water surfaces. For example, in a
film camera works the lens focuses an image onto the film surface. The color film
has three layers of emulsion, each layer being sensitive to a different color, and
the (slide) film records on each tiny spot of the film to reproduce the same color
as the image projected onto it, the same as the lens saw. This is an analog image,
the same as our eyes can see, so we can hold the developed film up and look at it.
Images may be two-dimensional, such as a photograph, or three-dimensional,
such as a statue or a hologram. An image in a broad sense also refers to any two-
dimensional figure such as a map, a graph, a pie chart, or an abstract painting. In
this sense, images can also be rendered manually, such as by drawing, painting, carv-
ing; can be rendered automatically by printing or computer graphics technology; or
can be developed by a combination of methods, especially in a pseudophotograph. In
photography, visual media, and the computer industries, the phrase “still image” re-
fers to a single static image that is distinguished from a kinetic or moving image
(often called video), which emphasizes that one is not talking about movies. The
phrase “still image” is often used in very precise or pedantic technical writing
such as an image compression standard.
In this book, we consider two-dimensional still images in a broad sense. Thus, an
analog image (physical image) I defined in the “real world” is considered to be a
function of two real variables as follows:
I ¼ fIðx; yÞ˛½0; Bj0  x  X; 0  y  Yg (1.1)
where I(x,y) is the amplitude (e.g., brightness or intensity) of the image at the real
coordinate position (x,y), B is the possible maximum amplitude, and X and Y define
the maximum coordinates. An image may be considered to contain subimages
Lossless Information Hiding in Images. http://dx.doi.org/10.1016/B978-0-12-812006-4.00001-2 1
Copyright © 2017 Zhejiang University Press Co., Ltd., published by Elsevier Inc. All rights reserved.
2 CHAPTER 1 Introduction

sometimes referred to as regions. This concept reflects the fact that images
frequently contain collections of objects, each of which can be the basis for a region.

1.1.1.2 Digital Images

A digital image is the numeric representation of an analog image (physical image).
Any image from a scanner, from a digital camera, or in a computer is a digital image.
Depending on whether the image resolution is fixed or not, it may be of vector or
raster type. Raster images are created through the process of scanning source
artwork or “painting” with a photo editing or paint program such as Corel Photo-
PAINT or Adobe PhotoShop. A raster image is a collection of dots called pixels.
Pixel is a computer word formed from PICture ELement, because a pixel is a tiny
colored square that is the smallest element of the digital image. Scanned images
and web images [Joint Photographic Experts Group (JPEG) and graphics inter-
change format (GIF) files] are the most common forms of raster images. Vector im-
ages are created through the process of drawing with vector illustration programs
such as CorelDRAW or Adobe Illustrator. The word “vector” is a synonym for
line. A vector image is a collection of connected lines and curves that produce ob-
jects. When creating a vector image in a vector illustration program, node or drawing
points are inserted and lines and curves connect the nodes together. Sometimes, both
raster and vector elements will be combined in one image, for example, in the case
of a billboard with text (vector) and photographs (raster). By itself, the term “digital
image” usually refers to raster images.
In this book, we mainly consider two-dimensional still raster images. Raster im-
ages can be created by a variety of input devices and techniques, such as digital cam-
eras, scanners, coordinate-measuring machines, seismographic profiling, and
airborne radar. A digital camera creates a digital picture with a charge-coupled de-
vice or complementary metal oxide semiconductor chip behind the lens. The lens
focuses the physical image onto the digital sensor, which is constructed with a
grid of many tiny light-sensitive cells, or sensors, arranged to divide the total picture
area into rows and columns composed of a huge number of very tiny subareas called
pixels, as shown in Fig. 1.1. Each sensor inspects and remembers the color of the tiny
area. A digital camera remembers the color by digitizing the analog color into three
digital values representing the color (i.e., three components, red, green, and blue,

FIGURE 1.1
Digitalization of the Physical Image by Pixels.
 1.1 Background 3

called RGB), or sometimes one digital value representing the brightness of the color.
Similarly, a scanner has a one-row array of similar cells, and a carriage motor moves
this row of sensors down the page, making columns in many rows to form the full
image grid. Both scanners and cameras generate images composed of pixels, and
a pixel contains the digital RGB color data or brightness of one tiny surface area.
This process is called digitization. Printers and video screens are digital devices
too, and their only purpose in life is to display pixels.
From these descriptions, we come to know that a digital image contains a fixed
number of rows and columns of pixels. Pixels are the smallest individual element in
a digital image, holding quantized values that represent the brightness of a given co-
lor at any specific point. Typically, the pixels are stored in computer memory as a
raster image or raster map, a two-dimensional array of small integers. Thus, a digital
image I can be defined as an array, or a matrix, of square pixels arranged in columns
and rows as follows:
I ¼ fIðm; nÞj0  m  M  1; 0  n  N  1g (1.2)
where I(m,n) is the color data or brightness value of the pixel at the mth column and
nth row, and M and N define the width (number of columns) and height (number of
rows) of the digital image.
According to the range of I(m,n), we can classify digital images into binary im-
ages, grayscale images, and color images, and three examples are shown in Fig. 1.2.
In color images, each pixel’s color sample has three numerical RGB components to
represent the color of that tiny area, i.e., I(m,n) is denoted by (R, G, B) with R˛[0,
255], G˛[0, 255], and B˛[0, 255]. Typically, for each pixel, its three RGB compo-
nents are three 8-bit numbers. These 3 bytes (1 byte for each RGB) compose a 24-bit
color. Each byte can have 256 possible values, ranging from 0 to 255. In the RGB
system, we know red and green make yellow. Thus (255, 255, 0) means both red
and green are fully saturated (255 is as large as an 8-bit value can be), with no

FIGURE 1.2
The (a) Binary, (b) Grayscale, and (c) Color Images of Lena.
Another random document with
no related content on Scribd:
and disjoin the closest affinities. And if we note the differences of
leaves, or even flowers, we fall into the same difficulty; for many
plants very different in kind have leaves very similar, as Polygonum
and Hypericum, Ernea and Sesamois, Apium and Ranunculus; and
plants of the same genus have sometimes very different 376 leaves,
as the several species of Ranunculus and of Lactuca. Nor will color
or shape of the flowers help us better; for what has Vitis in common
with Œnanthe, except the resemblance of the flower?” He then goes
on to say, that if we seek a too close coincidence of all the
characters we shall have no Species; and thus shows us that he had
clearly before his view the difficulty, which he had to attack, and
which it is his glory to have overcome, that of constructing Natural
Orders.
46 Lib. vi.

47 Lib. vii.

48 Lib. x.

49 Lib. xii.

50 Lib. xi.

51 Lib. i. cap. xii. p. 25.

But as the principles of Cæsalpinus are justified, on the one hand,

by their leading to Natural Orders, they are recommended on the
other by their producing a System which applies through the whole
extent of the vegetable kingdom. The parts from which he takes his
characters must occur in all flowering-plants, for all such plants have
seeds. And these seeds, if not very numerous for each flower, will be
of a certain definite number and orderly distribution. And thus every
plant will fall into one part or other of the same system.
 It is not difficult to point out, in this induction of Cæsalpinus, the
two elements which we have so often declared must occur in all
inductive processes; the exact acquaintance with facts, and the
general and applicable ideas by which these facts are brought
together. Cæsalpinus was no mere dealer in intellectual relations or
learned traditions, but a laborious and persevering collector of plants
and of botanical knowledge. “For many years,” he says in his
Dedication, “I have been pursuing my researches in various regions,
habitually visiting the places in which grew the various kinds of
herbs, shrubs, and trees; I have been assisted by the labors of many
friends, and by gardens established for the public benefit, and
containing foreign plants collected from the most remote regions.”
He here refers to the first garden directed to the public study of
Botany, which was that of Pisa, 52 instituted in 1543, by order of the
Grand Duke Cosmo the First. The management of it was confided
first to Lucas Ghini, and afterwards to Cæsalpinus. He had collected
also a herbarium of dried plants, which he calls the rudiment of his
work. “Tibi enim,” he says, in his dedication to Francis Medici, Grand
Duke of Etruria, “apud quem extat ejus rudimentum ex plantis libro
agglutinatis a me compositum.” And, throughout, he speaks with the
most familiar and vivid acquaintance of the various vegetables which
he describes.
52 Cuv. 187.

But Cæsalpinus also possessed fixed and general views

concerning the relation and functions of the parts of plants, and
ideas of symmetry 377 and system; without which, as we see in other
botanists of his and succeeding times, the mere accumulation of a
knowledge of details does not lead to any advance in science. We
have already mentioned his reference to general philosophical
principles, both of the Peripatetics and of his own. The first twelve
chapters of his work are employed in explaining the general structure
of plants, and especially that point to which he justly attaches so
much importance, the results of the different situation of the cor or
corculum of the seed. He shows 53 that if we take the root, or stem,
or leaves, or blossom, as our guide in classification, we shall
separate plants obviously alike, and approximate those which have
merely superficial resemblances. And thus we see that he had in his
mind ideas of fixed resemblance and symmetrical distribution, which
he sedulously endeavored to apply to plants; while his acquaintance
with the vegetable kingdom enabled him to see in what manner
these ideas were not, and in what manner they were, really
applicable.
53 Lib. i. cap. xii.

The great merit and originality of Cæsalpinus have been generally

allowed, by the best of the more modern writers on Botany. Linnæus
calls him one of the founders of the science; “Primus verus
systematicus;” 54 and, as if not satisfied with the expression of his
admiration in prose, hangs a poetical garland on the tomb of his
hero. The following distich concludes his remarks on this writer:

Quisquis hic extiterit primos concedet honores

Cæsalpine tibi; primaque serta dabit:

and similar language of praise has been applied to him by the best
botanists up to Cuvier, 55 who justly terms his book “a work of
genius.”
54 Philosoph. Bot. p. 19.

55 Cuv. Hist. 193.

 Perhaps the great advance made in this science by Cæsalpinus, is
most strongly shown by this; that no one appeared, to follow the path
which he had opened to system and symmetry, for nearly a century.
Moreover, when the progress of this branch of knowledge was
resumed, his next successor, Morison, did not choose to
acknowledge that he had borrowed so much from so old a writer;
and thus, hardly mentions his name, although he takes advantage of
his labors, and even transcribes his words without
acknowledgement, as I shall show. The pause between the great
invention of Cæsalpinus, and its natural sequel, the developement
and improvement of his method, is so marked, that I 378 will, in order
to avoid too great an interruption of chronological order, record some
of its circumstances in a separate section.

Sect. 3.—Stationary Interval.

The method of Cæsalpinus was not, at first, generally adopted. It

had, indeed, some disadvantages. Employed in drawing the
boundary-lines of the larger divisions of the vegetable kingdom, he
had omitted those smaller groups, Genera, which were both most
obvious to common botanists, and most convenient in the
description and comparison of plants. He had also neglected to give
the Synonyms of other authors for the plants spoken of by him; an
appendage to botanical descriptions, which the increase of botanical
information and botanical books had now rendered indispensable.
And thus it happened, that a work, which must always be considered
as forming a great epoch in the science to which it refers, was
probably little read, and in a short time could be treated as if it were
quite forgotten.
 In the mean time, the science was gradually improved in its
details. Clusius, or Charles de l’Ecluse, first taught botanists to
describe well. “Before him,” says Mirbel, 56 “the descriptions were
diffuse, obscure, indistinct; or else concise, incomplete, vague.
Clusius introduced exactitude, precision, neatness, elegance,
method: he says nothing superfluous; he omits nothing necessary.”
He travelled over great part of Europe, and published various works
on the more rare of the plants which he had seen. Among such
plants, we may note now one well known, the potato; which he
describes as being commonly used in Italy in 1586; 57 thus throwing
doubt, at least, on the opinion which ascribes the first introduction of
it into Europe to Sir Walter Raleigh, on his return from Virginia, about
the same period. As serving to illustrate, both this point, and the
descriptive style of Clusius, I quote, in a note, his description of the
flower of this plant. 58
56 Physiol. Veg. p. 525.

57 Clusius. Exotic. iv. c. 52, p. lxxix.

58 “Papas Peruanorum. Arachidna, Theoph. forte. Flores

elegantes, uncialis amplitudinis aut majores, angulosi, singulari
folio constantes, sed ita complicato ut quinque folia discreta
videantur, coloris exterius ex purpura candicantis, interius
purpurascentis, radiis quinque herbaceis ex umbilico stellæ instar
prodeuntibus, et totidem staminibus flavis in umbonem
coeuntibus.”
He says that the Italians do not know whence they had the
plant, and that they call it Taratouffli. The name Potato was, in
England, previously applied to the Sweet Potato (Convolvulus
batatas), which was the common Potato, in distinction to the
Virginian Potato, at the time of Gerard’s Herbal. (1597?) Gerard’s
figures of both plants are copied from those of Clusius.
It may be seen by the description of Arachidna, already quoted
 from Theophrastus, (above,) that there is little plausibility in
Clusius’s conjecture of the plant being known to the ancients. I
need not inform the botanist that this opinion is untenable.

379 The addition of exotic species to the number of known plants

was indeed going on rapidly during the interval which we are now
considering. Francis Hernandez, a Spaniard, who visited America
towards the end of the sixteenth century, collected and described
many plants of that country, some of which were afterwards
published by Recchi. 59 Barnabas Cobo, who went as a missionary to
America in 1596, also described plants. 60 The Dutch, among other
exertions which they made in their struggle with the tyranny of Spain,
sent out an expedition which, for a time, conquered the Brazils; and
among other fruits of this conquest, they published an account of the
natural history of the country. 61 To avoid interrupting the connexion
of such labors, I will here carry them on a little further in the order of
time. Paul Herman, of Halle, in Saxony, went to the Cape of Good
Hope and to Ceylon; and on his return, astonished the botanists of
Europe by the vast quantity of remarkable plants which he
introduced to their knowledge. 62 Rheede, the Dutch governor of
Malabar, ordered descriptions and drawings to be made of many
curious species, which were published in a large work in twelve folio
volumes. 63 Rumphe, another Dutch consul at Amboyna, 64 labored
with zeal and success upon the plants of the Moluccas. Some
species which occur in Madagascar figured in a description of that
island composed by the French Commandant Flacourt. 65 Shortly
afterwards, Engelbert Kæmpfer, 66 a Westphalian of great
acquirements and undaunted courage, visited Persia, Arabia Felix,
the Mogul Empire, Ceylon, Bengal, Sumatra, Java, Siam, Japan;
Wheler travelled in Greece and Asia Minor; and Sherard, the English
consul, published an account of the plants of the neighborhood of
Smyrna.
59 Nova Plantarum Regni Mexicana Historia, Rom. 1651, fol.

60 Sprengel, Gesch. der Botanik, ii. 62.

61 Historia Naturalis Brasiliæ, L. B. 1648, fol. (Piso and Maregraf).

62 Museum Zeylanicum, L. B. 1726.

63 Hortus Malabaricus, 1670–1703.

64 Herbarium Amboinense, Amsterdam, 1741–51, fol.

65 Histoire de la grande Isle Madagascar, Paris, 1661.

66 Amœnitates Exoticæ, Lemgov. 1712. 4to.

380 At the same time, the New World excited also the curiosity of
botanists. Hans Sloane collected the plants of Jamaica; John
Banister those of Virginia; William Vernon, also an Englishman, and
David Kriege, a Saxon, those of Maryland; two Frenchmen, Surian
and Father Plumier, those of Saint Domingo.

We may add that public botanical gardens were about this time
established all over Europe. We have already noticed the institution
of that of Pisa in 1543; the second was that of Padua in 1545; the
next, that of Florence in 1556; the fourth, that of Bologna, 1568; that
of Rome, in the Vatican, dates also from 1568.

The first transalpine garden of this kind arose at Leyden in 1577;

that of Leipzig in 1580. Henry the Fourth of France established one
at Montpellier in 1597. Several others were instituted in Germany;
but that of Paris did not begin to exist till 1626; that of Upsal,
afterwards so celebrated, took its rise in 1657, that of Amsterdam in
1684. Morison, whom we shall soon have to mention, calls himself,
in 1680, the first Director of the Botanical Garden at Oxford.

[2nd Ed.] [To what is above said of Botanical Gardens and

Botanical Writers, between the times of Cæsalpinus and Morison, I
may add a few circumstances. The first academical garden in France
was that at Montpellier, which was established by Peter Richier de
Belleval, at the end of the sixteenth century. About the same period,
rare flowers were cultivated at Paris, and pictures of them made, in
order to supply the embroiderers of the court-robes with new
patterns. Thus figures of the most beautiful flowers in the garden of
Peter Robins were published by the court-embroiderer Peter Vallet,
in 1608, under the title of Le Jardin du Roi Henry IV. But Robins’
works were of great service to botany; and his garden assisted the
studies of Renealmus (Paul Reneaulme), whose Specimen Historiæ
Plantarum (Paris, 1611), is highly spoken of by the best botanists.
Recently, Mr. Robert Brown has named after him a new genus of
Irideæ (Renealmia); adding, “Dixi in memoriam Pauli Renealmi,
botanici sui ævi accuratissimi, atque staminum primi scrutatoris; qui
non modo eorum numerum et situm, sed etiam filamentorum
proportionem passim descripsit, et characterem tetradynamicum
siliquosarum perspexit.” (Prodromus Floræ Novæ Hollandiæ, p.
448.)

The oldest Botanical Garden in England is that at Hampton Court,

founded by Queen Elizabeth, and much enriched by Charles II. and
William III. (Sprengel, Gesch. d. Bot. vol. ii. p. 96.)]

In the mean time, although there appeared no new system which

381 commanded the attention of the botanical world, the feeling of
the importance of the affinities of plants became continually more
strong and distinct.

Lobel, who was botanist to James the First, and who published his
Stirpium Adversaria Nova in 1571, brings together the natural
families of plants more distinctly than his predecessors, and even
distinguishes (as Cuvier states, 67 ) monocotyledonous from
dicotyledonous plants; one of the most comprehensive division-lines
of botany, of which succeeding times discovered the value more
completely. Fabius Columna, 68 in 1616, gave figures of the
fructification of plants on copper, as Gessner had before done on
wood. But the elder Bauhin (John), notwithstanding all that
Cæsalpinus had done, retrograded, in a work published in 1619, into
the less precise and scientific distinctions of—trees with nuts; with
berries; with acorns; with pods; creeping plants, gourds, &c.: and no
clear progress towards a system was anywhere visible among the
authors of this period.
67 Cuv. Leçons, &c. 198.

68 Ib. 206.

While this continued to be the case, and while the materials, thus
destitute of order, went on accumulating, it was inevitable that the
evils which Cæsalpinus had endeavored to remedy, should become
more and more grievous. “The nomenclature of the subject 69 was in
such disorder, it was so impossible to determine with certainty the
plants spoken of by preceding writers, that thirty or forty different
botanists had given to the same plant almost as many different
names. Bauhin called by one appellation, a species which Lobel or
Matheoli designated by another. There was an actual chaos, a
universal confusion, in which it was impossible for men to find their
way.” We can the better understand such a state of things, from
having, in our own time, seen another classificatory science,
Mineralogy, in the very condition thus described. For such a state of
confusion there is no remedy but the establishment of a true system
of classification; which by its real foundation renders a reason for the
place of each species; and which, by the fixity of its classes, affords
a basis for a standard nomenclature, as finally took place in Botany.
But before such a remedy is obtained, men naturally try to alleviate
the evil by tabulating the synonyms of different writers, as far as they
are able to do so. The task of constructing such a Synonymy of
botany at the period of which we speak, was undertaken by Gaspard
Bauhin, the brother of John, but nineteen years younger. This work,
the Pinax Theatri Botanici, was printed 382 at Basil in 1623. It was a
useful undertaking at the time; but the want of any genuine order in
the Pinax itself, rendered it impossible that it should be of great
permanent utility.
69 Ib. 212.

After this period, the progress of almost all the sciences became
languid for a while; and one reason of this interruption was, the wars
and troubles which prevailed over almost the whole of Europe. The
quarrels of Charles the First and his parliament, the civil wars and
the usurpation, in England; in France, the war of the League, the
stormy reign of Henry the Fourth, the civil wars of the minority of
Louis the Thirteenth, the war against the Protestants and the war of
the Fronde in the minority of Louis the Fourteenth; the bloody and
destructive Thirty Years’ War in Germany; the war of Spain with the
United Provinces and with Portugal;—all these dire agitations left
men neither leisure nor disposition to direct their best thoughts to the
promotion of science. The baser spirits were brutalized; the better
were occupied by high practical aims and struggles of their moral
nature. Amid such storms, the intellectual powers of man could not
work with their due calmness, nor his intellectual objects shine with
their proper lustre.

At length a period of greater tranquillity gleamed forth, and the

sciences soon expanded in the sunshine. Botany was not inert amid
this activity, and rapidly advanced in a new direction, that of
physiology; but before we speak of this portion of our subject, we
must complete what we have to say of it as a classificatory science.

Sect. 4.—Sequel to the Epoch of Cæsalpinus. Further Formation

and Adoption of Systematic Arrangement.

Soon after the period of which we now speak, that of the restoration
of the Stuarts to the throne of England, systematic arrangements of
plants appeared in great numbers; and in a manner such as to show
that the minds of botanists had gradually been ripening for this
improvement, through the influence of preceding writers, and the
growing acquaintance with plants. The person whose name is
usually placed first on this list, Robert Morison, appears to me to be
much less meritorious than many of those who published very
shortly after him; but I will give him the precedence in my narrative.
He was a Scotchman, who was wounded fighting on the royalist side
in the civil wars of England. On the triumph of the republicans, he
withdrew to France, when he became director of the garden of
Gaston, Duke of Orléans at Blois; and there he came under the
notice of our Charles 383 the Second; who, on his restoration,
summoned Morison to England, where he became Superintendent of
the Royal Gardens, and also of the Botanic Garden at Oxford. In
1669, he published Remarks on the Mistakes of the two Bauhins, in
which he proves that many plants in the Pinax are erroneously
placed, and shows considerable talent for appreciating natural
families and genera. His great systematic work appeared from the
University press at Oxford in 1680. It contains a system, but a
system, Cuvier says, 70 which approaches rather to a natural method
than to a rigorous distribution, like that of his predecessor
Cæsalpinus, or that of his successor Ray. Thus the herbaceous
plants are divided into climbers, leguminous, siliquose, unicapsalar,
bicapsular, tricapsular, quadricapsular, quinquecapsular; this division
being combined with characters derived from the number of petals.
But along with these numerical elements, are introduced others of a
loose and heterogeneous kind, for instance, the classification of
herbs as lactescent and emollient. It is not unreasonable to say, that
such a scheme shows no talent for constructing a complete system;
and that the most distinct part of it, that dependent on the fruit, was
probably borrowed from Cæsalpinus. That this is so, we have, I
think, strong proof; for though Morison nowhere, I believe, mentions
Cæsalpinus, except in one place in a loose enumeration of botanical
writers, 71 he must have made considerable use of his work. For he
has introduced into his own preface a passage copied literally 72 from
the dedication of Cæsalpinus; which passage we have already
quoted (p. 374,) beginning, “Since all science consists in the
collection of similar, and the distinction of dissimilar things.” And that
the mention of the original is not omitted by accident, appears from
this; that Morison appropriates also the conclusion of the passage,
which has a personal reference, “Conatus sum id præstare in
universa plantarum historia, ut si quid pro ingenii mei tenuitate in
hujusmodi studio profecerim, ad communem utilitatem proferrem.”
That Morison, thus, at so long an interval after the publication of the
work of Cæsalpinus, borrowed from him without acknowledgement,
and adopted his system so as to mutilate it, proves that he had
neither the temper nor the talent of a discoverer; and justifies us
withholding from him the credit which belongs to those, who, in his
time, resumed the great undertaking of constructing a vegetable
system.
70 Cuv. Leçons, &c. p. 486.

71 Pref. p. i.

72 Ib. p. ii.

Among those whose efforts in this way had the greatest and
earliest 384 influence, was undoubtedly our countryman, John Ray,
who was Fellow of Trinity College, Cambridge, at the same time with
Isaac Newton. But though Cuvier states 73 that Ray was the model of
the systematists during the whole of the eighteenth century, the
Germans claim a part of his merit for one of their countrymen,
Joachim Jung, of Lubeck, professor at Hamburg. 74 Concerning the
principles of this botanist, little was known during his life. But a
manuscript of his book was communicated 75 to Ray in 1660, and
from this time forwards, says Sprengel, there might be noticed in the
writings of Englishmen, those better and clearer views to which
Jung’s principles gave birth. Five years after the death of Jung, his
Doxoscopia Physica was published, in 1662; and in 1678, his
Isagoge Phytoscopica. But neither of these works was ever much
read; and even Linnæus, whom few things escaped which
concerned botany, had, in 1771, seen none of Jung’s works.
73 Leçons Hist. Sc. p. 487.

74 Sprengel, ii. 27.

 75 Ray acknowledges this in his Index Plant. Agri Cantab. p. 87,
and quotes from it the definition of caulis.

I here pass over Jung’s improvements of botanical language, and

speak only of those which he is asserted to have suggested in the
arrangement of plants. He examines, says Sprengel, 76 the value of
characters of species, which, he holds, must not be taken from the
thorns, nor from color, taste, smell, medicinal effects, time and place
of blossoming. He shows, in numerous examples, what plants must
be separated, though called by a common name, and what most be
united, though their names are several.
76 Sprengel, ii. 29.

I do not see in this much that interferes with the originality of Ray’s
method, 77 of which, in consequence of the importance ascribed to it
by Cuvier, as we have already seen, I shall give an account,
following that great naturalist. 78 I confine myself to the ordinary
plants, and omit the more obscure vegetables, as mushrooms,
mosses, ferns, and the like.
77 Methodus Plantarum Nova, 1682. Historia Plantarum, 1686.

78 Cuv. Leçons Hist. Sc. 488.

Such plants are composite or simple. The composite flowers are

those which contain many florets in the same calyx. 79 These are
subdivided according as they are composed altogether of complete
florets, 385 or of half florets, or of a centre of complete florets,
surrounded by a circumference or ray of demi-florets. Such are the
divisions of the corymbiferæ, or compositæ.
79 Involucrum, in modern terminology.
 In the simple flowers, the seeds are naked, or in a pericarp. Those
with naked seeds are arranged according to the number of the
seeds, which may be one, two, three, four, or more. If there is only
one, no subdivision is requisite: if there are two, Ray makes a
subdivision, according as the flower has five petals, or a continuous
corolla. Here we come to several natural families. Thus, the flowers
with two seeds and five petals are the Umbelliferous plants; the
monopetalous flowers with two seeds are the Stellatæ. He founds
the division of four-seeded flowers on the circumstance of the leaves
being opposite, or alternate; and thus again, we have the natural
families of Asperifoliæ, as Echium, &c., which have the leaves
alternate, and the Verticillatæ, as Salvia, in which the leaves are
opposite. When the flower has more than four seeds, he makes no
subdivision.

So much for simple flowers with naked seeds. In those where the
seeds are surrounded by a pericarp, or fruit, this fruit is large, soft,
and fleshy, and the plants are pomiferous; or it is small and juicy, and
the fruit is a berry, as a Gooseberry.

If the fruit is not juicy, but dry, it is multiple or simple. If it be simple,

we have the leguminose plants. If it be multiple, the form of the
flower is to be attended to. The flower may be monopetalous, or
tetrapetalous, or pentapetalous, or with still more divisions. The
monopetalous may be regular or irregular; so may the tetrapetalous.
The regular tetrapetalous flowers are, for example, the Cruciferæ, as
Stock and Cauliflower; the irregular, are the papilionaceous plants,
Peas, Beans, and Vetches; and thus we again come to natural
families. The remaining plants are divided in the same way, into
those with imperfect, and those with perfect, flowers. Those with
imperfect flowers are the Grasses, the Rushes (Junci), and the like;
among those with perfect flowers, are the Palmaceæ, and the
Liliaceæ.

We see that the division of plants is complete as a system; all

flowers must belong to one or other of the divisions. Fully to explain
the characters and further subdivisions of these families, would be to
write a treatise on botany; but it is easily seen that they exhaust the
subject as far as they go.

Thus Ray constructed his system partly on the fruit and partly on
the flower; or more properly, according to the expression of Linnæus,
386 comparing his earlier with his later system, he began by being a
fructicist, and ended by being a corollist. 80 ~Additional material in the
3rd edition.~
80 Ray was a most industrious herbalizer, and I cannot
understand on what ground Mirbel asserts (Physiol. Veg., tom. ii.
p. 531,) that he was better acquainted with books than with plants.

As we have said, a number of systems of arrangement of plants

were published about this time, some founded on the fruit, some on
the corolla, some on the calyx, and these employed in various ways.
Rivinus 81 (whose real name was Bachman,) classified by the flower
alone; instead of combining it with the fruit, as Ray had done. 82 He
had the further merit of being the first who rejected the old division,
of woody and herbaceous plants; a division which, though at
variance with any system founded upon the structure of the plants
was employed even by Tournefort, and only finally expelled by
Linnæus.
81 Cuv. Leçons, 491.

82 Historia Generalis ad rem Herbariam, 1690.

 It would throw little light upon the history of botany, especially for
our purpose, to dwell on the peculiarities of these transitory systems.
Linnæus, 83 after his manner, has given a classification of them.
Rivinus, as we have just seen, was a corollist, according to the
regularity and number of the petals; Hermann was a fructicist.
Christopher Knaut 84 adopted the system of Ray, but inverted the
order of its parts; Christian Knaut did nearly the same with regard to
that of Rivinus, taking number before regularity in the flower. 85
83 Philos. Bot. p. 21.

84 Enumeratio Plantarum, &c., 1687.

85 Linn.

Of the systems which prevailed previous to that of Linnæus,

Tournefort’s was by far the most generally accepted. Joseph Pitton
de Tournefort was of a noble family in Provence, and was appointed
professor at the Jardin du Roi in 1683. His well-known travels in the
Levant are interesting on other subjects, as well as botany. His
Institutio Rei Herbariæ, published in 1700, contains his method,
which is that of a corollist. He is guided by the regularity or
irregularity of the flowers, by their form, and by the situation of the
receptacle of the seeds below the calyx, or within it. Thus his classes
are—those in which the flowers are campaniform, or bell-shaped;
those in which they are infundibuliform, or funnel-shaped, as
Tobacco; then the irregular flowers, as the Personatæ, which
resemble an ancient mask; the Labiatæ, with their two lips; the
Cruciform; the Rosaceæ, with flowers like a rose; the Umbelliferæ;
the Caryophylleæ, as the 387 Pink; the Liliaceæ, with six petals, as
the Tulip, Narcissus, Hyacinth, Lily; the Papilionaceæ, which are
leguminous plants, the flower of which resembles a butterfly, as
Peas and Beans; and finally, the Anomalous, as Violet, Nasturtium,
and others.

Though this system was found to be attractive, as depending, in

an evident way, on the most conspicuous part of the plant, the
flower, it is easy to see that it was much less definite than systems
like that of Rivinus, Hermann, and Ray, which were governed by
number. But Tournefort succeeded in giving to the characters of
genera a degree of rigor never before attained, and abstracted them
in a separate form. We have already seen that the reception of
botanical Systems has depended much on their arrangement into
Genera.

Tournefort’s success was also much promoted by the author

inserting in his work a figure of a flower and fruit belonging to each
genus; and the figures, drawn by Aubriet, were of great merit. The
study of botany was thus rendered easy, for it could be learned by
turning over the leaves of a book. In spite of various defects, these
advantages gave this writer an ascendancy which lasted, from 1700,
when his book appeared, for more than half a century. For though
Linnæus began to publish in 1735, his method and his nomenclature
were not generally adopted till 1760.
 CHAPTER IV.

The Reform of Linnæus.

Sect. 1.—Introduction of the Reform.

A LTHOUGH, perhaps, no man of science ever exercised a

greater sway than Linnæus, or had more enthusiastic admirers,
the most intelligent botanists always speak of him, not as a great
discoverer, but as a judicious and strenuous Reformer. Indeed, in his
own lists of botanical writers, he places himself among the
“Reformatores;” and it is apparent that this is the nature of his real
claim to admiration; for the doctrine of the sexes of plants, even if he
had been the first to establish it, was a point of botanical physiology,
a province of the 388 science which no one would select as the
peculiar field of Linnæus’s glory; and the formation of a system of
arrangement on the basis of this doctrine, though attended with
many advantages, was not an improvement of any higher order than
those introduced by Ray and Tournefort. But as a Reformer of the
state of Natural History in his time, Linnæus was admirable for his
skill, and unparalleled in his success. And we have already seen, in
the instance of the reform of mineralogy, as attempted by Mohs and
Berzelius, that men of great talents and knowledge may fail in such
an undertaking.

It is, however, only by means of the knowledge which he displays,

and of the beauty and convenience of the improvements which he
proposes, that any one can acquire such an influence as to procure
his suggestions to be adopted. And even if original circumstances of
birth or position could invest any one with peculiar prerogatives and
powers in the republic of science, Karl Linné began his career with
no such advantages. His father was a poor curate in Smaland, a
province of Sweden; his boyhood was spent in poverty and privation;
it was with great difficulty that, at the age of twenty-one, he contrived
to subsist at the University of Upsal, whither a strong passion for
natural history had urged him. Here, however, he was so far
fortunate, that Olaus Rudbeck, the professor of botany, committed to
him the care of the Botanic Garden. 86 The perusal of the works of
Vaillant and Patrick Blair suggested to him the idea of an
arrangement of plants, formed upon the sexual organs, the stamens
and pistils; and of such an arrangement he published a sketch in
1731, at the age of twenty-four.
86 Sprengel, ii. 232.

But we must go forwards a few years in his life, to come to the

period to which his most important works belong. University and
family quarrels induced him to travel; and, after various changes of
scene, he was settled in Holland, as the curator of the splendid
botanical garden of George Clifford, an opulent banker. Here it
was 87 that he laid the foundation of his future greatness. In the two
years of his residence at Hartecamp, he published nine works. The
first, the Systema Naturæ, which contained a comprehensive sketch
of the whole domain of Natural History, excited general
astonishment, by the acuteness of the observations, the happy talent
of combination, and the clearness of the systematic views. Such a
work could not fail to procure considerable respect for its author. His
Hortus Cliffortiana 389 and Musa Cliffortiana added to this
impression. The weight which he had thus acquired, he proceeded to
use for the improvement of botany. His Fundamenta Botanica and
Bibliotheca Botanica appeared in 1736; his Critica Botanica and