Professional Documents
Culture Documents
Full Chapter Computer Vision and Image Analysis Digital Image Processing and Analysis 4Th Edition Scott E Umbaugh PDF
Full Chapter Computer Vision and Image Analysis Digital Image Processing and Analysis 4Th Edition Scott E Umbaugh PDF
https://textbookfull.com/product/digital-image-enhancement-
restoration-and-compression-digital-image-processing-and-
analysis-for-true-epub-scott-e-umbaugh/
https://textbookfull.com/product/feature-extraction-and-image-
processing-for-computer-vision-4th-edition-mark-nixon/
https://textbookfull.com/product/image-processing-and-computer-
vision-in-ios-oge-marques/
https://textbookfull.com/product/nonlinear-eigenproblems-in-
image-processing-and-computer-vision-guy-gilboa/
Computer Vision Pattern Recognition Image Processing
and Graphics Renu Rameshan
https://textbookfull.com/product/computer-vision-pattern-
recognition-image-processing-and-graphics-renu-rameshan/
https://textbookfull.com/product/image-processing-computer-
vision-and-pattern-recognition-hamid-r-arabnia/
https://textbookfull.com/product/feature-extraction-and-image-
processing-for-computer-vision-fourth-edition-aguado/
https://textbookfull.com/product/digital-image-watermarking-
theoretical-and-computational-advances-intelligent-signal-
processing-and-data-analysis-1st-edition-borra/
https://textbookfull.com/product/applied-computer-vision-and-
image-processing-proceedings-of-iccet-2020-volume-1-brijesh-iyer/
ISTUDY
Computer Vision and Image Analysis
Computer Vision and Image Analysis, focuses on techniques and methods for image analysis and their use in the
development of computer vison applications. The field is advancing at an ever increasing pace, with applications
ranging from medical diagnostics to space exploration. The diversity of applications is one of the driving forces
that make it such an exciting field to be involved in for the 21st century. This book presents a unique engineering
approach to the practice of computer vision and image analysis, which starts by presenting a global model to help
gain an understanding of the overall process, followed by a breakdown and explanation of each individual topic.
Topics are presented as they become necessary for understanding the practical imaging model under study, which
provides the reader with the motivation to learn about and use the tools and methods being explored.
The book includes chapters on image systems and software, image analysis, edge, line and shape detection,
image segmentation, feature extraction and pattern classification. Numerous examples, including over 500 color
images are used to illustrate the concepts discussed. Readers can explore their own application development with
any programming language, including C/C ++, MATLAB®, Python, and R, and software is provided for both the
Windows/C/C ++ and MATLAB environment.
The book can be used by the academic community in teaching and research, with over 700 PowerPoint Slides and
a complete Solutions Manual to the over 150 included problems. It can also be used for self-study by those involved
with developing computer vision applications, whether they are engineers, scientists, or artists. The new edition
has been extensively updated and includes numerous problems and programming exercises that will help the
reader and student develop their skills.
Computer Vision and
Image Analysis
Digital Image Processing and Analysis
Fourth Edition
Scott E Umbaugh
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the accuracy of the text or exer-
cises in this book. This book’s use or discussion of MATLAB® software or related products does not constitute endorsement or sponsorship by
The MathWorks of a particular pedagogical approach or particular use of the MATLAB® software.
Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity
of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced
in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not
been acknowledged please write and let us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any
electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming and recording, or in any information
storage or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, access www.copyright.com or contact the Copyright Clearance Center,
Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. For works that are not available on CCC please contact mpkbookspermissions@
tandf.co.uk
Trademark notice: Product or corporate names may be trademarks or registered trademarks and are used only for identification and explanation
without intent to infringe.
DOI: 10.1201/9781003221135
Typeset in Palatino
by codeMantra
To Robin and David
and
In loving memory of
Digital image processing and analysis is a field that continues to experience rapid growth, with applications ranging
from areas such as space exploration to the entertainment industry. The diversity of applications is one of the driving
forces that make it such an exciting field to be involved in for the 21st century. Digital image processing can be defined as
the acquisition and processing of visual information by computer. This book presents a unique approach to the prac-
tice of digital image processing, and it will be of interest to those who want to learn about and use computer imaging
techniques.
Digital image processing can be divided into two primary application areas, human vision and computer vision,
with image analysis tying these two together. Computer Vision and Image Analysis, focuses on the tools needed for
development of computer vision applications. The culminating chapter provides the reader with examples of using the
tools discussed throughout the book necessary for developing computer vision applications. The automatic identifica-
tion of land types in satellite images, robotic control of a mars rover and the automatic classification of abnormalities
in medical images are all examples of computer vision applications. Human vision applications involve manipula-
tion of image data for viewing by people. Examples include the development of better compression algorithms, spe-
cial effects imaging for motion pictures and the restoration of satellite images distorted by atmospheric disturbance.
eResources (PowerPoint slides, software documentation, Solutions manuals, and quizzes) are also available on the
Routledge website: www.routledge.com\9781032071299
development tools and moved it earlier in the book – it is now Chapter 2 and was Chapter 11 in the previous edi-
tion. The creation of a GUI for the new CVIP MATLAB Toolbox will provide an additional platform for imaging
exploration and development, which is of particular interest to engineers.
The CVIP MATLAB Toolbox is a major new part of the continuing development of the comprehensive CVIPtools envi-
ronment. In conjunction with the previous development tools, the Computer Vision and Image Processing Algorithm
Test and Analysis Tool (CVIP-ATAT) and the CVIP Feature Extraction and Pattern Classification Tool (CVIP-FEPC), we
now realize a much more powerful development environment. The new Windows version of CVIPtools, which has
been integrated even more throughout the book, in conjunction with the new development tools, creates a valuable
environment for learning about imaging as well as providing a set of reusable tools for applications development.
Approach
To help motivate the reader, I have taken an approach to learning that presents topics as needed. This approach
starts by presenting a global model to help gain an understanding of the overall process, followed by a breakdown
and explanation of each individual topic. Instead of presenting techniques or mathematical tools when they fit into
a nice, neat theoretical framework, topics are presented as they become necessary for understanding the practical
imaging model under study. This approach provides the reader with the motivation to learn about and use the tools
and topics, because they see an immediate need for them. For example, the mathematical process of convolution
is introduced when it is needed for an image zoom algorithm, and morphological operations are introduced when
morphological filtering operations are needed after image segmentation. This approach also makes the book more
useful as a reference, or for those who may not work through the book sequentially, but will reference a specific
section as the need arises.
and representation. Chapter 2 includes material about the CVIPtools software development environment, both C/
C++ and MATLAB, to ease the reader into programming at an early stage. Chapter 3 leads the reader through the
various stages of image analysis, culminating in a simple binary computer vision application example. Chapter
4 discusses edge, line, boundary and shape detection. Chapter 5 provides an overview of image segmentation
methods, with representative examples of each algorithm type, and shows how morphological filtering is used as
segmentation postprocessing to clarify object shapes. Chapter 5 also discusses how segmentation and morphologi-
cal filters are applied to computer vision system development. Chapter 6 includes feature selection, feature extrac-
tion and feature analysis. To complete the tools needed for image analysis, and our view of computer vision as
deployed image analysis systems, Chapter 7 covers algorithm development and pattern classification methods. The
final chapter, Chapter 8, has more detailed examples of application development, including the Feature Extraction
and Pattern Classification (FEPC) and Algorithm Test and Analysis Tool (ATAT) development tools, as well as
new edited conference papers describing computer vision applications developed with CVIPtools. The reader who
wants to develop applications within a CVIP-development environment will find what is needed here. Also note
that complete documentation for the software is available on the publisher’s web site for the book: www.routledge.
com/9781032071299 as “Online Docs”.
CVIPlab program and the associated laboratory exercises. With the addition of the new GUI for the CVIP Toolbox for
MATLAB, the users can explore the functions in a manner similar to using the Windows GUI. The CVIPlab program
and the CVIP MATLAB Toolbox both allow the user access to any of the already-defined CVIPtools functions, rang-
ing from general purpose input/output and matrix functions to more advanced transform functions and complex
imaging algorithms; some of these functions are state-of-the-art algorithms since CVIPtools is continually being
enhanced at the Computer Vision and Image Processing Laboratory at Southern Illinois University Edwardsville
(SIUE).
MATLAB® is a registered trademark of The MathWorks, Inc. For product information, please contact:
The MathWorks, Inc.
3 Apple Hill Drive
Natick, MA 01760-2098 USA
Tel: 508-647-7000
Fax: 508-647-7001
E-mail: info@mathworks.com
Web: www.mathworks.com
Acknowledgments
I thank Southern Illinois University Edwardsville, specifically the School of Engineering and the Electrical and
Computer Engineering Department, for their support of this endeavor. I also thank all the students who have taken
my imaging courses and provided valuable feedback regarding the learning and teaching of computer vision and
image analysis.
The initial version of the CVIPtools software was developed primarily by myself and a few graduate students:
Gregory Hance, Arve Kjoelen, Kun Luo, Mark Zuke and Yansheng Wei; without their hard work and dedication,
the foundation that was built upon for the subsequent versions would not be solid. The next major Windows ver-
sion of CVIPtools was developed primarily by myself and Iris Cheng, Xiaohe Chen, Dejun Zhang and Huashi Ding.
Additional students who contributed were Husain Kagalwalla and Sushma Gouravaram.
The next version of CVIPtools was initially developed by Patrick Solt and Evan Bian. The work was completed by
Patrick Solt, Jhansi Akkineni, Mouinika Mamidi, Pelin Guvenc, Serkan Kefel and Hari Krishna Akkineni. Jhansi
Akkineni served as project coordinator, and her efforts deserve special recognition. Her dedication to the project and
the help she provided to others working on the project were substantial. Iris Cheng, who was the primary contributor on
the original Windows version of CVIPtools, also deserves special recognition for her continued support as a consultant
on the project. CVIPtools development continued with Hari Siva Kumar Reddy Bhogala, followed by Sai Chaitanya
Gorantla. The current version of CVIPtools was created by Rohini Dahal, Gita Pant, John M. Hilbing (Jack) and Charles
Stacey. All the graduate students who worked on the CVIPtools project deserve my special recognition and thanks.
The Computer Vision and Image Processing Algorithm Test and Analysis Tool, CVIP-ATAT, underwent many
changes during its development. The initial version was provided by Sid Smith and Jeremy Wood. Geer Shaung
and Evan Bian provided substantial new development for this tool, which helped to make it more practical for its
use with CVIPtools. Further development was completed, tested and utilized in projects by Pelin Guvenc. Hari
Siva Kumar Reddy Bhogala worked on the improvement of CVIP-ATAT and Ravneet Kaur performed testing for
the previous version. The CVIP-ATAT was worked on extensively by Charles Stacey, who has provided wonderful
improvements and updated methods for the software.
The Computer Vision and Image Processing Feature Extraction and Pattern Classification Tool, CVIP-FEPC, was
created and developed by Patrick Solt, and its development was partially funded by the Long Island Veterinary
Specialists. Additional work was performed for the improvement and enhancement of CVIP-FEPC by Kumari
Heema Poudel. Following that, Rohini Dahal took over programming development with Norsang Lama providing
extensive testing to create the first robust version of CVIP-FEPC. The current version has undergone enhancements,
primarily implemented by Charles Stacey. All the hard work and dedication of my graduate students is greatly
appreciated.
The MATLAB CVIP Toolbox was initially developed by the porting of CVIPtools C functions to mex (MATLAB
executable) files by Krishna Regmi, whose great effort laid the foundation and provided valuable insight into
our future development. The mex file development for the project was continued by Deependra Mishra. After
encountering numerous complications with various versions of compilers and operating systems, we decided to
continue development of the CVIP Toolbox using m-files. Mehrdad Alvandipour and Norsang Lama teamed up
with Deependra to complete the first version consisting primarily of m-files. Lakshmi Gorantla continued devel-
opment and testing before the release of version 2.5. Their long hours and perseverance during development is
greatly appreciated. This first version of the MATLAB CVIP Toolbox has gotten extensive use and has been a benefit
to those who are using CVIPtools and want to perform development in the MATLAB environment. For the new
version of the MATLAB Toolbox, we added a GUI and a number of functions. Julian Buritica was instrumental
with the GUI development, with the work continued by Hridoy Biswas and Al Mahmud. This version of the CVIP
Toolbox for MATLAB will be released in conjunction with this new book edition.
In small but important parts of CVIPtools, public domain software was used, and kudos to those who pro-
vided it: Jef Pokanzer’s pbmplus, Sam Leffler’s TIFF library, Paul Heckbert’s Graphics Gems, the Independent JPEG
Group’s software, Yuval Fisher’s fractal code and the Texas Agricultural Experiment Station’s code for the original
texture features.
I’d like to thank those who contributed photographs and images: Howard Ash, Mark Zuke, Mike Wilson,
Tony Berke, George Dean, Sara Sawyer, Sue Eder, Jeff Zuke, Bill White, the National Oceanic and Atmospheric
Administration, NASA and MIT. H. Kipsang Choge, Charles Stacey and Hridoy Biswas deserve credit for helping
xviii Acknowledgments
out with some of the figures, and I thank them for this work. Thanks also to David, Robin, Michael, Jeanie, Jeff,
Mom, Greg, Glicer, Gaby, Dad, Pat, Angi, Tyler, Connor, Kayla, Aaron, Ava, Kenley, Kaiden, Chris, Ryder, Chad,
Jamie, Noah, Rohini, Gita, Hridoy, Charles, Iris and Alice for letting me use their smiling faces in some of the
figures.
I also want to extend my gratitude to my graduate students and others who helped with the review of various
sections of the book. This primarily includes Julian Buritica, Charles Stacey, Hridoy Biswas and Al Mahmud, as
well as the numerous students who have used my books to learn about imaging in the courses I teach and those
taught world-wide. I also want to thank all the other authors and coauthors of the papers used in Chapter 8.
I also thank the publisher, Nora Konopka at CRC Press/Taylor & Francis Group, for having the vision, sagacity
and good taste in publishing the fourth edition of the book. Nora has been very supportive throughout the project
and very helpful in making sure that everything was put together the way I wanted. The entire staff at the CRC
Press, especially Prachi Mishra, Glenon Butler and Swapnil Joshi, have been very helpful in getting and keeping
the project rolling, and special thanks go to them. Their encouragement and enthusiasm is much appreciated; they
all have done a wonderful job managing project details as we approach production. Thanks also go to Gayathri
Tamilselvan and the production staff who survived my many requests regarding the book layout, but still managed
to get it done in a timely manner.
Finally, I thank my family for all their contributions; without them, this book would not have been possible. I
thank my mom who instilled in me a love of learning and a sense of wonder about the world around me; my dad,
who taught me how to think like an engineer and the importance of self-discipline and hard work. I want to thank
my brothers, Greg and Jeff, for being there during those formative years. And I am especially grateful to Jeanie,
Robin, David and Michael, who lived through the ordeal and endured the many long hours I put into the new
edition of this book.
Author
Dr. Scott E Umbaugh is a University Distinguished Research Professor of Electrical and Computer Engineering
and Graduate Program director for the Department of Electrical and Computer Engineering at Southern Illinois
University Edwardsville (SIUE). He is also the director of the Computer Vision and Image Processing (CVIP)
Laboratory at SIUE. He has been teaching computer vision and image processing, as well as computer and electri-
cal engineering design, for over 30 years. His professional interests include computer vision and image processing
education, research and development of both human and computer vision applications, and engineering design
education.
Prior to his academic career, Dr. Umbaugh worked as a computer design engineer and project manager in the avi-
onics and telephony industries. He has been a computer imaging consultant since 1986 and has provided consulting
services for the aerospace, medical and manufacturing industries with projects ranging from automatic identification
of defects in microdisplay chips to analysis of thermographic images for clinical diagnosis of brain disease. He has
performed research and development for projects funded by the National Institutes of Health, the National Science
Foundation and the U.S. Department of Defense.
Dr. Umbaugh is the author or co-author of numerous technical papers, two edited books and multiple editions
of his textbooks on computer vision and image processing. His books are used at academic and research organiza-
tions throughout the world. He has served on editorial boards and as a reviewer for a variety of IEEE journals, and
has evaluated research monographs and textbooks in the imaging field.
Dr. Umbaugh received his B.S.E. degree with honors from Southern Illinois University Edwardsville in 1982,
M.S.E.E. in 1987 and Ph.D. in 1990 from the Missouri University of Science and Technology, where he was a
Chancellor’s Fellow. He is a life senior member of the Institute of Electrical and Electronic Engineers (IEEE) and a
member of Sigma Xi and the International Society for Optical Engineering (SPIE). Dr. Umbaugh is also the primary
developer of the CVIPtools software package and the associated CVIP MATLAB Toolbox.
1
Digital Image Processing and Analysis
1.1 Introduction
Digital image processing is a field that continues to grow, with new applications being developed at an ever-increas-
ing pace. It is a fascinating and exciting area to be involved in today with application areas ranging from medicine
to the entertainment industry to the space program. The internet, with its ease of use via web browsers, combined
with the advances in computer power and network bandwidth has brought the world into our offices and into our
homes. One of the most interesting aspects of this information revolution is the ability to send and receive com-
plex data that transcends ordinary written text. Visual information, transmitted in the form of digital images, has
become a major method of communication for the 21st century. With the advent of modern mobile phones, this
information is available anywhere and anytime, with uses limited only by the imagination.
Digital image processing can be defined as the acquisition and processing of visual information by computer.
The importance of digital image processing is derived from the fact that our primary sense is our visual sense.
Our vision system allows us to gather information without the need for physical interaction; it enables us to
analyze many types of information directly from pictures or video. It provides us with the ability to navigate
about our environment, and the human visual system is the most sophisticated, advanced neural system in
the human body. Most of the scientific discoveries and advancements have relied on the visual system for their
development – from the discovery of fire to the design of a smart phone.
The information that can be conveyed in images has been known throughout the centuries to be extraordi-
nary – one picture is worth a thousand words. Fortunately, this is the case because the computer representation
of an image requires the equivalent of many thousands of words of data, and without a corresponding amount of
information, the medium would be prohibitively inefficient. The massive amount of data required for images is a
primary reason for the development of many subareas within the field of image processing, such as image segmen-
tation and image compression. Another important aspect of computer imaging involves the ultimate “receiver” of
the visual information – in some cases the human visual system and in others the computer itself.
This distinction allows us to separate digital image processing into two primary application areas: (1) com-
puter vision applications and (2) human vision applications, with image analysis being a key component in the
development and deployment of both (Figure 1.1-1). In computer vision applications, the processed (output) images
COMPUTER HUMAN
IMAGE
VISION VISION
ANALYSIS
APPLICATIONS APPLICATIONS
FIGURE 1.1-1
Digital image processing can be separated into computer vision and human vision applications, with image analysis being
essential to both.
DOI: 10.1201/9781003221135-1
2 Computer Vision and Image Analysis
are for use by a computer, while in human vision applications, the output images are for human consumption. The
human visual system and the computer as a vision system have varying limitations and strengths, and the image
processing specialist needs to be aware of the functionality of these two very different systems. The human
vision system is limited to visible wavelengths of light, which cover only a small portion of the electromagnetic
(EM) spectrum. The computer is capable of dealing with the entire EM spectrum, ranging from gamma rays to
radio waves, and it can process other imaging modalities such as ultrasound and magnetic resonance imaging
(MRI).
Historically, the field of digital image processing grew from electrical engineering as an extension of the signal
processing branch, while the computer science discipline was largely responsible for developments in computer
vision applications. At some universities, these two are still separate and distinct, but the commonalities and the
perceived needs have brought the two together. Here, we will simply refer to digital image processing as the gen-
eral field while allowing for separate application areas in computer and human vision. As shown in Figure 1.1-1,
image analysis applies to both applications’ areas.
Image analysis involves the examination of the image data to facilitate solving an imaging problem. Image analysis
methods comprise the major components of a computer vision system, where the system is to analyze images and
have a computer act on the results. In one sense, a computer vision application is simply a deployed image analysis
system. In the development of a human vision image processing application, many images must be examined and
tested so image analysis is necessary during the development of the system.
This book focuses on computer vision and image analysis. Chapters 1 and 2 introduce the basic concepts, and will
provide the necessary background for those who are new to the field, and an overview of software tools that have
been developed for use with the book. This includes a discussion of image acquisition, imaging systems and image
representation. Chapters 3–7 cover a system model for the image analysis process and then describe each major part
of this model in separate chapters. Each of these chapters concludes with a key points section, followed by references
and suggestions for further reading, and a series of exercises to help the learning process. The exercises include
definitions, discussions, problems, advanced problems and computer exercises using the CVIPtools software and
programming exercises. Chapter 8 concludes the book with a discussion of application development with an over-
view of the more advanced tools in the development environment, the Algorithm Test and Analysis Tool and the
Feature Pattern and Classification Tool.
For the programming exercises, the reader has multiple options. The new CVIP MATLAB Toolbox is available
to MATLAB users, and the CVIPlab prototype program can be used by C, C++ or C# programmers. The program-
ming environment provided with the CVIPtools software is a comprehensive environment for computer vision
and image processing education and application development. Chapter 8 describes the development tools in more
detail, including examples, and provides applications that were developed with the CVIPtools environment. For
downloading, installing and updating CVIPtools, and a description of the CVIPtools software organization, see the
web site: cviptools.siue.edu. Note that more information is available on the publisher’s web site, including function
quick reference lists, application papers and useful image processing resources.
types of environments – from manufacturing plants to hospital surgical suites to the surface of Mars. For example,
in manufacturing systems, computer vision is often used for quality control. There, the computer vision system will
scan manufactured items for defects and provide control signals to a robotic manipulator to automatically remove
defective parts. To develop an application of this nature, an image database consisting of sample images is first
created. Next, image analysis is applied to develop the necessary algorithms to solve the problem. One interesting
example of this type of system involves the automated inspection of microdisplay chips.
Microdisplay chips are used in digital cameras, projection systems, televisions, heads-up-displays and any appli-
cation that requires a small imaging device. Prior to the design of this computer vision system, these chips were
inspected manually – a process that is slow and prone to error. Once the market demand for these chips accelerated,
the manual inspection process was not practical. In Figure 1.2-1, we see the microdisplay chip inspection system along
with two sample images. The original images were captured at 3:1 magnification, which means each picture element
in the microdisplay chip corresponds to a 3 × 3 array in the image. The system automatically finds various types of
defects in the chips, such as the pixel defects and the faint display defects shown here.
An active area of computer vison research today involves applications for facial recognition. This is desired for
many reasons, including recognizing terrorist suspects in airport crowds or pictures of your long-lost cousin in a
stack of photos. The first step in facial recognition is to find faces in an image. An algorithm to select subimages
within an image that are likely to contain faces was developed using convolutional neural networks (see Chapters 5
and 7), and a resulting output image is shown in Figure 1.2-2a. In this image, the subimages that contain faces have
been marked with boxes. In addition to facial recognition, Figure 1.2-2b shows results from an algorithm developed
to identify chairs and computer monitors.
Another interesting computer vision application that required image analysis for algorithm development
involved the automatic counting and grading of lumber. Before this system was implemented, this task was done
manually, which was a boring task, had an unacceptable error rate and was inefficient. This application was chal-
lenging due to the variation in the lumber stack, such as gaps in between boards, variation in the wood color,
cracks in the boards or holes in the boards. Figure 1.2-3 shows an example of the system in operation, with a
sample input image, processed image and system output. The processed image is used by high-level software
to count and grade the lumber in stack. With the system in place, the lumberyard can minimize errors, increase
efficiency and provide their workers with more rewarding tasks.
Image analysis is used in the development of many computer vision applications for the medical community,
with the only certainty being that the variety of applications will continue to grow. Current examples of medical
systems being developed include systems to automatically diagnose skin lesions, systems to aid neurosurgeons
during brain surgery and systems to automatically perform clinical tests. Systems that automate the diagnostic pro-
cess are being developed for various reasons, including pre-screening to determine appropriate subsequent tests or
analysis, as tools by medical professionals where specialists are unavailable, or to act as consultants to the primary
care givers, and they may serve their most useful purpose in the training of medical professionals. These types of
systems may be experimental, and some are still in the research stages, but more and more of them are being inte-
grated into medical practice. Computer vision systems that are being used in the surgical suite have already been
used to improve the surgeon’s ability to “see” what is happening in the body during surgery, and consequently,
improve the quality of medical care. Systems are also currently being used for tissue and cell analysis; for example,
to automate applications that require the identification and counting of certain types of cells (see Figure 1.2-4).
Computer vision is being used more and more in medical applications, but it will still be some time before we see
computers playing doctor like the holographic doctor in the Star Trek series.
The field of law enforcement and security is an active area for image analysis research and development, with
applications ranging from automatic identification of fingerprints to DNA analysis. Security systems to identify
people by retinal scans, facial scans and the veins in the hand have been developed. Reflected ultraviolet (UV)
imaging systems are being used to find latent fingerprints, shoeprints, body fluids and bite marks that are not vis-
ible to the human visual system. Infrared (IR) imaging to count bugs has been used at Disney World to help keep
their greenery green. Currently, cars are equipped with systems to automatically avoid accidents, and systems are
under development and being tested to fully automate our transportation systems to make travel safer. The United
States space program and the Defense department, with their needs for robots with visual capabilities, are actively
involved in image analysis research and development. Applications range from autonomous vehicles to target
tracking and identification. Satellites orbiting the earth collect massive amounts of image data every day, and these
images are automatically scanned to aid in making maps, predicting the weather and helping us to understand the
changes taking place on our home planet.
4 Computer Vision and Image Analysis
(a)
FIGURE 1.2-1
Microdisplay chip inspection. (a) Computer vision system for microdisplay chip inspection, (b) microdisplay chip image1 at
3:1 magnification, (c) Image1 after pixel defect detection, (d) Image1 after blob analysis to find faint defects, (e) microdisplay chip
Image2 at 3:1 magnification, (f) Image2 after pixel defect detection and (g) Image2 after blob analysis to find faint defects (photo
courtesy of Mike Wilson and Iris Cheng, Westar Display Technologies, Inc.)
Digital Image Processing and Analysis 5
(a)
(b)
FIGURE 1.2-2
Face and object recognition. (a) Image after being processed by a computer vision algorithm to find and identify faces. (b)
Image processed to identify chairs and monitors.
6 Computer Vision and Image Analysis
(a) (b)
(c)
FIGURE 1.2-3
Lumber grading. (a) Image captured by the computer vision system for lumber counting grading, (b) intermediate image after
processing, (c) example of system output after software analysis. (Photo courtesy of Tony Berke, River City Software.)
FIGURE 1.2-4
Necrotic liver tissue classification. An image that was used to develop a computer vision algorithm to automatically segment
and classify necrotic (dead) liver tissue from surrounding tissue. The research investigated the toxicity of an acetaminophen
overdose, the leading cause of acute liver failure in the United States, with the goal of identifying novel therapeutic targets. The
small gray areas are blood vessels and its area should be subtracted from the necrotic area. (Image courtesy of David Umbaugh
and the University of Kansas Medical Center.)
Phone
Monitor
Computer System
Camera
Printer
Phone
Image Sensor
Flash Drive
FIGURE 1.3-1
Digital image processing system hardware.
8 Computer Vision and Image Analysis
any desired analysis or processing on the image data. Additionally, we may also use software to control the image
acquisition and storage process.
The computer system may be a general-purpose computer with an imaging device connected. Images may be
acquired by the computer via a standalone camera or phone camera, or they can be input from any image sensing
device or downloaded from the internet. The primary device for capturing live images is the camera, which can
use either a digital or analog format. Digital cameras store image data in a format similar to that used directly by
the computer, while analog cameras output a continuous video signal which must be modified to create an image
suitable for computer processing. Although digital cameras are the newer format, analog cameras are still used
in many applications due to the large installed base, well-established standards and inexpensive, easily available
parts.
A standard analog video camera requires a frame grabber, or image digitizer, to interface with the computer.
The frame grabber is a special-purpose piece of hardware that accepts a standard analog video signal and outputs
an image in the form that a computer can understand – a digital image. Analog video standards vary throughout
the world; RS-170A, RS-330 and RS-343A are the monochrome video standards in the North America and Japan.
RS-343A is used for high-resolution video with 675–1,023 lines per frame. CCIR is the monochrome standard used
primarily in Europe. The three color video standards are NTSC, PAL and SECAM. NTSC is used in North America,
Japan and parts of South America, while PAL is used in parts of Africa, Asia and Northern Europe. SECAM is used
in parts of Africa, Asia, Eastern Europe, Russia and France. NTSC is 525 lines, 30 frames (60 fields) per second, 2:1
interlaced standard. PAL and SECAM are 625 lines, 25 frames (50 fields) per second, 2:1 interlaced standards.
The process of transforming a standard analog video signal into a digital image is called digitization. This trans-
formation is necessary because the standard video signal is in analog (continuous) form, and the computer requires
a digitized or sampled version of that continuous signal. A typical video signal contains frames of video informa-
tion, where each frame corresponds to a full screen of visual information. Each frame may then be divided into
fields, and each field consists of alternating lines of video information. In Figure 1.3-2a, we see the typical image on
a display device, where the solid lines represent one field of information and the dotted lines represent the other
field. These two fields make up one frame of visual information. This two fields per frame model is referred to as
interlaced video. Some types of video signals, called progressive scan or non-interlaced video, have only one field per
frame. Non-interlaced video is typically used in computer monitors.
In Figure 1.3-2b, we see the electrical signal that corresponds to one line of video information. Note the horizontal
synch pulse between each line of information, and this synchronization pulse tells the display hardware to start a
new line. After one frame has been displayed, a longer synchronization pulse, called the vertical synch pulse, tells
the display hardware to start a new field or frame.
The analog video signal is converted to a digital image by sampling the continuous signal at a fixed rate. In
Figure 1.3-3, we see one line of a video signal being sampled (digitized) by instantaneously measuring the volt-
age of the signal at fixed intervals in time. The value of the voltage at each instant is converted into a number that
is stored, corresponding to the brightness of the image at that point. Note that the image brightness at a point
(a) (b)
Horizontal
sync pulse
Time
FIGURE 1.3-2
The video signal. (a) One frame, two fields. Each field is obtained by taking very other line. One field is represented by solid
lines and the other by dotted lines. (b) The analog video signal. One line of information is one row of image data, and the hori-
zontal synch pulse is the control signal for the hardware to designate the end of a line.
Another random document with
no related content on Scribd:
The Project Gutenberg eBook of Gratitude
This ebook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this ebook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.
Title: Gratitude
Language: English
The first of a new series of stories about the Competents in Alaska. The
season ended, they throw their last shovelfuls of dirt and bid good-by to
the homely cabin on the mountainside. But their last days on the claim
were not to pass without adventure.