Advanced Digital Image Processing and Its Applications in Big Data

Ankur Dumka, Alaknanda Ashok, Parag Verma, Poonam Verma

• Publisher: CRC Press
• Year: 2020
• eBook ISBN: 9781000281392
• Edition: 1
• Topic: Computer Science
• Subtopic: Computer Graphics
• Index: Computer Science

This book covers the technology of digital image processing in various fields with big data and their applications. Readers will understand various technologies and strategies used in digital image processing as well as handling big data, using machine-learning techniques. This book will help to improve the skills of students and researchers in such fields as engineering, agriculture, and medical imaging.

There is a need to be able to understand and analyse the latest developments of digital image technology. As such, this book will cover:

· Applications such as biomedical science and biometric image processing, content-based image retrieval, remote sensing, pattern recognition, shape and texture analysis

· New concepts in color interpolation to produce the full color from the sub-pattern bare pattern color prevalent in today's digital cameras and other imaging devices

· Image compression standards that are needed to serve diverse applications

· Applications of remote sensing, medical science, traffic management, education, innovation, and analysis in agricultural design and image processing

· Both soft and hard computing approaches at great length in relation to major image processing tasks

· The direction and development of current and future research in many areas of image processing

· A comprehensive bibliography for additional research (integrated within the framework of the book)

This book focuses not only on theoretical and practical knowledge in the field but also on the traditional and latest tools and techniques adopted in image processing and data science. It also provides an indispensable guide to a wide range of basic and advanced techniques in the fields of image processing and data science.

Part I

Concept and Background of Image Processing, Techniques, and Big Data

1Introduction to Advanced Digital Image Processing

1.1INTRODUCTION

Digital image processing is being used ubiquitously in various fields, which has been increasing exponentially. This is majorly due to extensive use of digital images in the fields of remote sensing, medicine, machine vision, video processing, microscopic imaging, and so on. Image processing requires manipulation of image data using various electronic devices and softwares. Along with devices, digital image processing requires the application of different algorithms as per the requirements to convert a physical image into a digital image to fetch desired information or features (Figure 1.1).

A digital image is a representation of two-dimensional images as a finite set of digital picture elements termed as pixels. These pixel values represent various parameters like gray levels, height, colors, opacities, etc. of an image in the form of binary digits, and the binary digits can be represented in the form of mathematical equations. The digital image size can be determined by the matrix used to store the pixels based on their size. In order to access a particular pixel in the digital image, the relevant coordinates at x and y axis are defined. Each pixel has its own unique intensity and brightness. Pixels in an image will have different values as per an image or else the images may not appear different from each other. Various mixtures of colors will produce a color image. Pixel dimensions are the horizontal and the vertical measurements of an image. Each pixel is defined using the bit depth which is determined by the number of bits. Resolution is the spatial scale of digital images, is the indicator of the spatial frequency with which the images have been sampled, and can be measured in lpi, ppi, and dpi. Lpi stands for lines per inch and is used generally for magazine printing. Ppi stands for pixels per inch and refers to the pixel arrays depicting the real-world image. Dpi stands for dots per inch and is used to describe the printer’s resolution.

FIGURE 1.1Process of digital image processing.

1.2CATEGORIZATION OF DIGITAL IMAGES

Digital images may be categorized as shown in Figure 1.2.

FIGURE 1.2Categorization of digital images.

1.2.1Binary Image

Binary images are images whose pixels have only two possible values normally as 0 ad 1 displayed as black and white. This image is termed as monochrome or bitonal. Binary images are represented by pixels that can represent only one shade where each pixel consists of 1 bit each. It is an image that is composed exclusively of shades of only one color with the varying range from the brightest to the darkest hues.

1.2.2Black and White Image

This image consists of only two colors, i.e black and white color and is termed as black and white image. It combines black and white in a continuous fashion creating different ranges of gray. The color range is represented in 256 different gray values. These different shades lie between 0 and 255, where 0 refers to black, 255 refers to white (also known as panchromatic images), and intermediate shades refer to the neutral tonal values of black and white, which are commonly termed as grayscale image, whereas 127 stands for gray. Previous versions of monitors used 4 bits that could display only 16 shades of color between white and black. However, in present-day scenario, 8-bit grayscale is used to indicate that only 8 bits are used to store different shades of gray, thereby permitting 256 different intensities of black and white in each pixel.

1.2.38-Bit Color Format

This image format consists of 256 different shades of colors. This method requires storing the image information in the computer’s memory where each pixel is represented by using one byte, that is, 8 bits. Thus, maximum numbers of colors that can be displayed are not more than 256.

Color image is basically formed by three colors red, blue, and green to represent a coloured image. There are basically two forms of the 8-bit color graphics. One form utilizes each of the 256 entries for the red, green, and blue color thereby forming shades of 16,777,216 colors. In this approach, each 8 bit out of 24 bits describes the shades of red, green, and blue. Sometimes 18 bits or 12 bits can be used to define the shades of the color where 18 bits utilizes 6 bits for red, green, blue (RGB) forming a palette of 262,144 colors and 12 bits utilizes 4 bits for each RGB thereby forming a color palette of 4096 colors.

The other form of the 8-bit color format is three bits for red, three bits for green, and two bits for blue. This second form is often called 8-bit truecolor, as it does not use a palette at all. Most 8-bit image formats store a local image palette of 256 colors, as the graphics hardware’s global color palette will be overwritten with the local image palette, due to which it is highly possible to have distorted colors of the images. This is one of the major reasons that the 8-bit hardware programs are written along with the web browsers to be able to display images from various sources; each image may consist of its own palette, which will be finally mapped to a local palette, thereby causing some form of dithering. The popular file formats that consist of 8-bit formats are GIF, PNG, and BMP. In case the 24-bit image is converted into the 8-bit image, the image loses its quality and sharpness.

1.2.416 Color Format

In this type of image format, there are 65,536 types of different colors, and hence, it is termed as high color format. The 16-bit format is divided into three primary colors of red, green, and blue, and the distribution of the RGB can be 5 bits for red, 6 bits for green color, and 5 bits for representing the blue color. Generally, the distribution is like the above stated, and one extra bit is allocated to the green color, as it is soothing to the eyes among the three colors.

1.2.524-Bit Format

24-bit color format is also known as true color format. Like 16-bit color format, in a 24-bit color format, the 24 bits are again distributed in three different formats of red, green, and blue. Since 24 bits are equally divided on 8, they have been distributed equally between three different color channels. Their distributions are as follows:

• 8 bits for R (red),
• 8 bits for G (green),
• 8 bits for B (blue).

Compared to indexed color images, true color images lack a color lookup table. A pixel does not have an index referring to a specific color in the color lookup table. Every pixel has its own (RGB) color value and, depending on the file format, may also consist of a value for transparency (RGBA). The main ...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Preface
• Acknowledgments
• Authors
• Part I: Concept and Background of Image Processing, Techniques, and Big Data
• Part II: Advanced Image Processing Technical Phases for Big Data Analysis
• Part III: Various Application of Image Processing
• Index