Digital Image Processing

Module - 1

What Is Digital Image Processing?

The field of digital image processing refers to processing digital images by means of a digital
computer. Note that a digital image is composed of a finite number of elements, each of which has a
particular location and value. These elements are called picture elements, image elements, pels, and
pixels. Pixel is the term used most widely to denote the elements of a digital image.

Image Definition: An image is a two-dimensional function, f(x, y), where x and y are spatial
coordinates, and the amplitude at any point is called the intensity or gray level.

Applications of Digital Image Processing: Digital image processing has a broad range of
applications, extending across the entire electromagnetic spectrum, from gamma to radio waves.
Unlike humans, who perceive only the visual band, imaging machines can process images from
sources like ultrasound, electron microscopy, and computer-generated images.

Overlapping Fields: There’s no clear distinction between image processing, image analysis, and
computer vision. Image processing typically involves both inputs and outputs as images, while
computer vision aims to emulate human vision and intelligence, often using AI. Image analysis sits
between these, involving tasks like segmentation and object recognition.

Processing Levels: The continuum from image processing to computer vision can be categorized as
follows:

• Low-level processing: Basic operations like noise reduction, contrast enhancement, and
sharpening, where inputs and outputs are images.
• Mid-level processing: Involves tasks like segmentation, description, and classification, where
inputs are images, and outputs are attributes like edges and contours.
• High-level processing: Involves making sense of recognized objects and performing cognitive
functions.
The Origins of Digital Image Processing

Early Applications:
Digital images were first used in the newspaper industry in the 1920s, with the Bartlane cable system
transmitting pictures between London and New York. This reduced the transmission time from over a
week to less than three hours with an image of 5 levels of Gray. One of the image transmitted is as
shown in figure below.

Technological Improvements:
Early challenges involved improving image quality through better printing techniques. By 1929, the
Bartlane system could encode 15 levels of gray, enhancing image reproduction. One of the image is
as shown below.

Dependence on Computers:
Digital image processing truly began with the development of computers in the 1940s. The
introduction of key concepts like stored memory and conditional branching by John von Neumann,
along with technological advances like the transistor, integrated circuits, and microprocessors, paved
the way for digital image processing.
First Image Processing Tasks:
In 1964, NASA's Jet Propulsion Laboratory used computers to improve space images from Ranger 7,
marking the beginning of meaningful image processing tasks. This led to advancements in image
enhancement techniques for space exploration. Its image is shown below.

Medical Imaging Breakthrough:

The invention of Computerized Axial Tomography (CAT/CT) in the 1970s revolutionized medical
diagnostics by producing detailed cross-sectional images of the body.

Expanding Applications:
By the 1960s and 1970s, digital image processing extended to fields like medical imaging, remote
sensing, astronomy, and more. Techniques like image enhancement, restoration, and machine
perception became widely used in various domains, including industry, defense, and science.

Machine Perception:
Machine vision uses image processing techniques for automated tasks like character recognition,
product inspection, and military applications. These tasks often rely on information such as statistical
moments or Fourier transform coefficients, which may not resemble human interpretation of images.

Continued Growth:
With advancements in computing power, networking, and data transmission, digital image processing
continues to grow, finding applications in fields as diverse as archeology, law enforcement,
environmental monitoring, and space exploration.

Examples of Fields that Use Digital Image Processing

Digital image processing applications are diverse, and organizing them by image source helps in
understanding the field. The primary source of images is the electromagnetic (EM) energy spectrum,
which includes familiar images from X-rays and visible light. Electromagnetic waves can be seen as
either sinusoidal waves of varying wavelengths or as streams of massless particles called photons,
each carrying a bundle of energy. The EM spectrum ranges from high-energy gamma rays to low-
energy radio waves as shown.
GAMMA-RAY IMAGING
Gamma-ray imaging has significant applications in nuclear medicine and astronomy:

1. Nuclear Medicine:
o Involves injecting patients with a radioactive isotope that emits gamma rays.
o Bone Scan: Used to detect bone pathology (e.g., tumors or infections) by capturing
gamma-r ay emissions with detectors, as shown in Figure below.

o Positron Emission Tomography (PET): A patient is injected with a radioactive isotope

emitting positrons, which collide with electrons, producing gamma rays. This helps create
tomographic images, like Figure below, showing brain and lung tumors.
 2. Astronomy:
o Cygnus Loop: Figure below shows gamma radiation from the Cygnus Loop, a gas
cloud resulting from a supernova, using natural radiation.

Nuclear Reactor: Figure below depicts gamma radiation from a valve in a reactor, highlighting
areas of strong radiation.

X-RAY IMAGING
X-rays, among the oldest forms of EM radiation used for imaging, have applications in medicine,
industry, and astronomy:

1. Medical Imaging:
o Chest X-ray: Generated by placing the patient between an X-ray source and film,
showing areas where X-rays are absorbed differently, as in Fig. below,
 o Angiography: A catheter is used to inject a contrast medium to highlight blood
vessels for clearer imaging, as seen in Figure.

o CAT Scans: Provide high-resolution 3D images, like Figure below, by capturing

multiple slices of the body for detailed internal views.

2. Industrial Applications:
o X-rays: Used to inspect electronic circuit boards for defects, shown in Figure.
Industrial CAT scans can examine larger objects like plastic assemblies or rocket
motors.
 3. Astronomy:
o X-ray imaging is also used to capture celestial objects, like the Cygnus Loop,
in Figure below, revealing cosmic phenomena.

IMAGING IN THE ULTRAVIOLET BAND

Ultraviolet (UV) light has a wide range of applications, including lithography, industrial inspection,
microscopy, lasers, biological imaging, and astronomy. Two key areas where UV imaging is prevalent
are microscopy and astronomy:

1. Fluorescence Microscopy: UV light is essential in fluorescence microscopy, a rapidly

advancing field. Fluorescence occurs when UV light excites electrons in certain materials,
causing them to emit visible light. The fluorescence microscope uses excitation light to
irradiate a specimen and captures the weaker fluorescent light emitted, which stands out against
a dark background. This method is ideal for studying both naturally fluorescing materials
(primary fluorescence) and those treated with chemicals that fluoresce (secondary
fluorescence). For example:
o Figure below shows a fluorescence microscope image of healthy corn.
 o Figure below depicts corn infected with "smut," a disease caused by parasitic fungi
that severely affects crops like corn.

2. Astronomy: UV imaging is also employed in astronomy, as shown in Figure below, which

captures the Cygnus Loop in the high-energy UV band.

IMAGING IN THE VISIBLE AND INFRARED BANDS

The visual band of the electromagnetic spectrum, often combined with infrared imaging, is the most
widely used for various applications. Key areas include:

1. Microscopy: Light microscopy has applications ranging from pharmaceuticals to materials

characterization. Processes like enhancement and measurement can be applied to these images.
2. Remote Sensing: Multispectral imaging, using visual and infrared bands, is used for
environmental monitoring. NASA’s LANDSAT satellite captures images to track population
growth, pollution, and other factors. Weather observation, such as monitoring storms like
Hurricane Katrina, is also a major application.
3. Infrared Imaging: The Night time Lights of the World project uses infrared imaging to map
human settlements, gas flares, and fires, helping estimate global energy consumption.
4. Industrial Inspection: Automated visual inspections of manufactured goods are common.
Examples include inspecting circuit boards for missing parts, checking pill containers for
 completeness, ensuring bottles are filled to an acceptable level, and detecting defects in plastic
parts and cereal.
5. Law Enforcement: Fingerprint analysis, currency tracking, and automated license plate
reading are widely used in security and surveillance systems, leveraging the visual band for
detailed image processing tasks.

Each of these examples illustrates the vast potential of digital image processing in the visual and
infrared spectrum across different fields.

IMAGING IN THE MICROWAVE BAND-

Microwave Imaging - Radar Applications:

• Radar Imaging: The dominant use of microwave imaging is radar, which can capture data in
any weather and lighting conditions, making it invaluable for regions where traditional
imaging fails. Some radar waves can penetrate clouds, vegetation, ice, and dry sand,
providing a unique way to explore otherwise inaccessible areas.
• Functionality: Imaging radar operates similarly to a flash camera but uses microwave pulses
instead of light for illumination. It captures the energy reflected back toward the radar
antenna, which is processed to create an image.
• Advantages: Radar can provide clear and detailed images regardless of weather conditions or
time of day. It’s used for mapping rugged terrains, monitoring agricultural regions, and
studying remote areas.
• Example: A spaceborne radar image of southeast Tibet illustrates the technology’s capacity
to capture detailed images of mountainous terrain and valleys, unaffected by clouds or
atmospheric disturbances. This allows for detailed study of regions like the Lhasa River
valley.

Radar's ability to bypass traditional barriers like cloud cover makes it a powerful tool for
environmental and geographical studies.

IMAGING IN THE RADIO BAND

Imaging in the Radio Band - Key Applications:

• Medical Use - MRI (Magnetic Resonance Imaging):

o MRI uses radio waves to generate detailed images of the body.
o Patients are placed in a powerful magnetic field, and radio pulses are passed through
the body.
o The pulses cause tissues to emit responding radio waves, which are captured and
analyzed by a computer to produce two-dimensional images.
o MRI is versatile, providing images in any plane of the body.
o Example: MRI images of the human knee and spine are commonly used for medical
diagnosis.
• Astronomical Use:
o Radio waves are used in astronomy to study celestial objects like pulsars.
o Example: The Crab Pulsar has been imaged in the radio band, showing a unique
perspective compared to images captured in other EM spectrum bands.
Fundamental Steps in Digital Image Processing

Figure – Fundamental steps in digital image processing.

1. Image Acquisition:
o First step in image processing; involves capturing or receiving an image in digital form.
o Pre-processing may include tasks like scaling or noise reduction.
2. Image Enhancement:
o Improves image quality for specific applications.
o Enhancement is subjective, depending on the problem (e.g., X-rays vs. satellite images).
o Introduced in early chapters for newcomers to understand processing techniques.
3. Image Restoration:
o Focuses on correcting image degradation based on mathematical models.
o Unlike enhancement, which is subjective, restoration is objective.
4. Color Image Processing:
o Essential due to the rise of digital images on the internet.
o Concepts of color models and basic color processing are covered in Chapter 6.
5. Wavelets and multiresolution processing:
o Wavelets represent images at various resolutions.
o Image compression reduces storage and bandwidth needs (e.g., JPEG format).
6. Compression:
o Image compression reduces storage and bandwidth needs (e.g., JPEG format).
7. Morphological Processing:
o Deals with extracting image components for shape representation and description.
8. Segmentation:
o Divides an image into parts or objects for analysis.
o Critical for tasks like object recognition; weak segmentation leads to failure.
 9. Representation and Description:
o Converts raw pixel data into a form suitable for computer processing.
o Focuses on external (boundary) or internal (region) characteristics of objects.
10. Recognition:
o Assigns labels to objects based on extracted features (e.g., “vehicle”).
11. Knowledge Base:
o Encodes domain knowledge into the system to guide processing.
o Controls interaction between processing modules.

Components of an Image Processing System

Figure - Components of a general-purpose image processing system.

1. Image Sensing and Digitization:

o Image acquisition requires a sensing device to detect energy (e.g., light) from the object
and a digitizer to convert this into digital data.
o Example: A digital video camera converts light into electrical signals, which are then
digitized.
2. Specialized Image Processing Hardware:
o Includes digitizers and Arithmetic Logic Units (ALUs) for fast data throughput, e.g.,
real-time averaging of images to reduce noise.
o Known as front-end subsystems due to their speed in handling tasks like video
processing at 30 frames/s.
 3. General-Purpose Computers:
o Most image processing systems use general-purpose computers ranging from PCs to
supercomputers.
o Custom computers may be used for specialized tasks, but general-purpose systems are
common for offline tasks.
4. Software:
o Software consists of specialized modules for specific tasks, often with the ability to
write custom code.
o Advanced packages allow integration of modules with general-purpose programming
languages.
5. Storage Requirements:
o Images require significant storage space, especially when uncompressed (e.g., 1 MB
for a 1024x1024 pixel image).
o Storage types include:

• Short-term: Computer memory or frame buffers.

• On-line: Magnetic disks or optical media for frequent access.
• Archival: Magnetic tapes or optical disks for massive, infrequently accessed
data.

6. Image Displays:
o Color flat-screen monitors are commonly used, driven by image display cards.
o Specialized applications may require stereo displays embedded in goggles.
7. Hardcopy Devices:
o Include laser printers, film cameras, inkjet printers, and digital storage like optical disks
and CD-ROMs.
o Film offers the highest resolution, while paper is preferred for written materials.
8. Networking:
o Networking is critical for image transmission, with bandwidth being a major concern,
especially over the Internet.
o Advances in broadband technology (e.g., optical fiber) are improving transmission
efficiency.

Elements of Visual Perception

Structure of the Human Eye

Figure – Simplified diagram of a cross section of the human eye.

The human eye is nearly spherical, with a diameter of approximately 20 mm. It is enclosed by three
membranes: the outer cornea and sclera, the choroid, and the retina.

The cornea is a transparent tissue that covers the eye's front, while the sclera is an opaque membrane
covering the rest of the eye. The choroid, located beneath the sclera, contains blood vessels essential
for eye nourishment and is heavily pigmented to reduce light scatter.

At the front, the choroid splits into the ciliary body and the iris, which controls the amount of light
entering the eye through the pupil. The lens, suspended by fibers from the ciliary body, is made of
fibrous cells and contains water and protein. It slightly filters visible light and protects against infrared
and ultraviolet light, which can damage the eye. Excessive clouding of the lens (cataracts) can impair
vision.

The retina, the innermost membrane, lines the back portion of the eye. It contains photoreceptors,
which are crucial for vision. Light entering the eye is focused on the retina, where two types of
receptors cones and rods are distributed. Cones (6-7 million) are concentrated in the fovea at the
retina’s center, enabling high-resolution, color vision in bright light (photopic vision). Rods (75-150
million) are distributed over the retina and provide overall, low-resolution vision, especially in low
light (scotopic vision). While rods help detect shapes in dim light, they are not sensitive to color, which
explains why objects lose their color in low-light conditions.

Distribution of rods and cones in the retina-

Figure – cones and rods distribution

The human eye contains two types of photoreceptors: rods and cones. There are approximately 75 to
150 million rods distributed over the retinal surface, which are responsible for low-light vision
(scotopic vision). Rods are not sensitive to color and provide a general view of the field, but their
connection to multiple nerve ends reduces the detail they can discern. This is why objects seen under
low light, like moonlight, appear colorless and less detailed.

Cones, on the other hand, number between 6 to 7 million and are highly concentrated in the fovea, the
central region of the retina, measuring about 1.5 mm in diameter. Cones are responsible for high-
resolution, color vision (photopic vision) and are most dense in the center of the fovea. As the distance
from the fovea increases, rod density rises, peaking at 20° off-axis, then decreases toward the periphery
of the retina.
The absence of receptors where the optic nerve exits the eye creates a blind spot. The fovea can be
considered roughly as a square sensor array of 1.5 mm by 1.5 mm, containing about 337,000 cones,
with a density of 150,000 elements per mm². Comparatively, modern CCD imaging sensors of medium
resolution can have a similar number of elements in arrays as small as 5 mm by 5 mm. Despite the
superficial nature of such comparisons, the resolving power of the human eye is quite similar to that
of current electronic imaging sensors.

Image Formation in the Eye

In an ordinary photographic camera, focusing is achieved by adjusting the distance between the lens
and the imaging plane, where the film or digital imaging chip is located. In contrast, the human eye
maintains a fixed distance between the lens and the retina, with the focal length adjusted by altering
the lens's shape. The ciliary body fibers facilitate this change, flattening the lens for distant objects and
thickening it for nearby objects. The distance from the lens center to the retina along the visual axis is
approximately 17 mm, with the eye's focal lengths ranging from about 14 mm to 17 mm. The longest
focal length occurs when the eye is relaxed and focused on objects more than 3 meters away.

To illustrate image formation, consider a scenario where a person views a 15-meter-high tree from a
distance of 100 meters. The geometry involved shows that the height of the tree in the retinal image
(denoted as hh) can be calculated using the ratio 15100=h1710015=17h, resulting
in h≈2.55h≈2.55 mm. The focused retinal image primarily occurs in the fovea, where perception is
based on the relative excitation of light receptors. These receptors transform radiant energy into
electrical impulses, which are then decoded by the brain for visual perception.

Brightness Adaptation and Discrimination

Brightness Adaptation-
Digital images are composed of discrete intensity levels, making the eye's ability to differentiate
between these levels crucial for effectively presenting image processing results. The human visual
system can adapt to a vast range of light intensities, approximately from the scotopic threshold to the
glare limit, spanning about 1010 levels. Research shows that subjective brightness, or the intensity
perceived by humans, follows a logarithmic relationship with the light intensity entering the eye.

Figure above illustrates this relationship, with the long solid curve indicating the intensity range the
visual system can adapt to. In photopic vision, the adaptable range is about 106. The transition between
scotopic (dim light) and photopic (bright light) vision occurs gradually, approximately between 0.001
to 0.1 millilambert, as shown by the double branches of the adaptation curve.

Brightness Discrimination-
The ability of the eye to discriminate between changes in light intensity at any specific adaptation
level is brightness discrimination.

A Classic Experiment -
• Having a subject look at a flat, uniformly illuminated area (diffuser-device that distribute
light evenly) large enough to occupy the entire field of view.
• It is illuminated from behind by a light source whose intensity, can be varied.
• To this field, an incremental illumination is added in the form of a short-duration flash that
appears as a circle in the centre of the uniformly illuminated field, as shown in figure.

• If DI is not bright enough, subject says “no”, indicating no perceivable changes

• If gets stronger, subject says “yes”, indicating a perceivable change.
• The Quantity DIc/I, where DIc is the increment of illumination discriminable 50% of the time
with background illumination, is called Weber ration.
• A small value of weber ratio means that a small percentage change in intensity is required.
This represents “good” brightness discrimination.
• A Large value of weber ratio means that a large percentage change in intensity is required.
This represents “poor” brightness discrimination.
Curve shows that brightness discrimination is poor (the Weber ratio is large) at low levels of
illumination, and it improves significantly (the Weber ratio decreases) as background illumination
increases. The two branches in the curve reflect the fact that at low levels of illumination vision is
carried out by the rods, whereas at high levels (showing better discrimination) vision is the function of
cones.
Two key phenomena demonstrate that perceived brightness is not directly correlated to light intensity.
The first phenomenon, called Machbands, illustrates how the visual system overshoots or undershoots
around the boundaries of regions with different intensities. In Figure below, the intensity of the stripes
is constant, but the eye perceives a scalloped brightness pattern near the boundaries.

The second phenomenon, known as simultaneous contrast, highlights that a region’s perceived
brightness depends on its surrounding areas, not just its own intensity. For example, in Figure below,
all center squares have the same intensity but appear darker as the background lightens. A common
experience of this phenomenon occurs when a piece of paper looks white on a desk but can appear
black when held up against a bright sky.

Additionally, optical illusions showcase how the human visual system can misinterpret visual
information. Figure below presents several such illusions:

• A square is perceived even though no defining lines are present.

• Few lines create the illusion of a full circle.
• Two horizontal line segments of the same length, yet one appears shorter.
• 45° lines appear far from parallel due to crosshatching, despite being equidistant and parallel.
Image Sensing and Acquisition
Image Acquisition Using a Single Sensor
A single sensor system, as shown in Figure , consists of a sensor like a photodiode, typically made
of silicon, where the output voltage is proportional to the light intensity. The sensor's selectivity can
be enhanced by using a filter in front of it. For instance, a green pass filter allows primarily green
light to pass through, thus amplifying the sensor's response to green light relative to other colors in
the visible spectrum.

To create a 2-D image using a single sensor, there must be relative movement between the sensor and
the area being imaged in both the x- and y-directions. Figure below illustrates one approach, in which
a film negative is placed on a rotating drum, providing displacement in one dimension. Meanwhile,
the single sensor is mounted on a lead screw that moves perpendicularly to the drum's motion,
enabling scanning in the other dimension. This arrangement allows for high-precision scanning, and
though it is slow, it is a cost-effective way to achieve high-resolution images due to the precise
mechanical control.
Image Acquisition Using Sensor Strips
A more common geometry for imaging than single sensors involves an in-line sensor strip, as shown
in Figure. This strip provides imaging elements in one direction, and motion perpendicular to the strip
completes the imaging in the second dimension. This configuration is widely used in flatbed
scanners and airborne imaging systems. In these applications, the imaging strip captures one line of
the image, while the movement of the object or the sensor completes the two-dimensional image.

In medical and industrial imaging, sensor strips are arranged in a ring configuration to
obtain cross-sectional (slice) images of 3D objects, as depicted in Figure. In this setup, a rotating X-
ray source provides illumination, and sensors opposite the source collect the energy that passes
through the object. This forms the basis for computerized axial tomography (CAT). However, the
data from the sensors require extensive processing via reconstruction algorithms to transform them
into meaningful cross-sectional images. As the object moves perpendicularly to the sensor ring,
multiple cross-sectional images are captured and stacked to form a 3D digital volume.
Image Acquisition Using Sensor Arrays

Figure above illustrates individual sensors arranged in a 2D array format, which is commonly used
in digital cameras and other electromagnetic or ultrasonic sensing devices. A well-known example is
the CCD (Charge-Coupled Device)array, widely utilized for its low-noise properties in applications
such as astronomy. CCD arrays can have configurations like 4000x4000 elements and more, making
them highly effective for high-resolution image capture.
The key feature of a 2D sensor array is that it can capture a complete image in one exposure, as the
energy pattern is projected onto the surface of the array. This eliminates the need for motion or
scanning, unlike the single sensor or sensor strip arrangements discussed previously.

Figure below outlines the operation of an array sensor system:

1. Illumination energy (such as light) reflects from a scene element and is collected by an
imaging system.
2. An optical lens focuses the incoming energy onto a focal plane, where the 2D sensor array is
located.
3. The sensor array produces outputs proportional to the amount of light energy received at each
sensor.
4. Analog circuitry sweeps these outputs and converts them into an analog signal, which is
then digitized by the imaging system, resulting in a digital image.

A Simple Image Formation Model

When an image is generated from a physical process, its intensity values are proportional to energy
radiated by a physical source (e.g., electromagnetic waves). As a consequence, f(x,y) must be nonzero
and finite; that is,
The function f(x, y) may be characterized by two components:
(1) the amount of source illumination incident on the scene being viewed, and
(2) the amount of illumination reflected by the objects in the scene. Appropriately, these are called the
illumination and reflectance components and are denoted by i(x, y) and r(x, y), respectively. The two
functions combine as a product to form f(x,y) :

Image Sampling and Quantization

To create a digital image, we need to convert the continuous sensed data into digital form. This
involves two processes: sampling and quantization.

Basic Concepts in Sampling and Quantization

To convert a continuous image into digital form, two key processes are
required: sampling and quantization. Sampling involves digitizing the x- and y-coordinates of the
image, while quantization digitizes the amplitude values.

• Sampling: This process captures discrete points from a continuous image along both the x-
and y-axes. As illustrated in Figure above, the amplitude (intensity) values of the image are
plotted along a line (AB). Sampling occurs by taking measurements at equally spaced intervals,
as shown by the small white squares superimposed on the function in Figure. These spatial
locations give rise to a discrete set of coordinate values, but the amplitude remains continuous.
• Quantization: To fully digitize the image, the amplitude values must also be converted into
discrete levels. Figure shows that the intensity range is divided into discrete intervals (in this
case, eight levels from black to white). Each sampled amplitude is assigned a value based on
its proximity to these intervals. This process is called quantization.
 • Final Digital Image: Once both sampling and quantization are completed, the image consists
of discrete values for both the coordinates and amplitude, as shown in Figure. Repeating this
process line by line for the entire image results in a two-dimensional digital image.

Representing Digital Images

Three basic ways to represent an image f(x,y)

1. Image plotted as a surface – here, two axis determines spatial location and third axis being
values of f as a function of the two spatial variables x and y as shown below.

2. Image displayed as a visual intensity array – here, intensity of each point is proportional to the
value of f at that point as shown below.

3. Image can be represented as 2D numerical array (0, 0.5, 1) representing dark, gray and white-
here, we have to display the numerical value of f(x,y) as an array(matrix) as shown below.
Mathematically, In equation form, we write the representation of an M x N numerical array as

Both sides of this equation are equivalent ways of expressing a digital image quantitatively. The right
side is a matrix of real numbers. Each element of this matrix is called an image element, picture
element, pixel, or pel.
It is advantageous to use a more traditional matrix notation to denote a digital image and its elements:

Where M is the number of rows and N is the number of columns. There are no restrictions on choosing
the values of M and N other than they have to be positive integers. However, due to storage and
quantizing hardware considerations, the number of intensity levels typically is an integer power of 2.
i.e L = 2K.

Sometimes, the range of values spanned by the gray scale is referred to informally as the dynamic
range. This is a term used in different ways in different fields. Here, we define the dynamic range of
an imaging system to be the ratio of the maximum measurable intensity to the minimum detectable
intensity level in the system. As a rule, the upper limit is determined by saturation and the lower limit
by noise Closely associated with this concept is image contrast, which we define as the difference in
intensity between the highest and lowest intensity levels in an image. When an appreciable number of
pixels in an image have a high dynamic range, we can expect the image to have high contrast.
Conversely, an image with low dynamic range typically has a dull, washed-out gray look.
The number, b, of bits required to store a digitized image is
Table below shows the number of bits required to store square images with various values of N and k.
The number of intensity levels corresponding to each value of k is shown in parentheses. When an
image can have 2k intensity levels, it is common practice to refer to the image as a “k-bit image.” For
example, an image with 256 possible discrete intensity values is called an 8-bit image. Note that storage
requirements for 8-bit images of size 1024 * 1024 and higher are not insignificant.

Spatial and Intensity Resolution

SPATIAL RESOLUTION à due to sampling

Spatial resolution is a measure of smallest discernible detail in an image.
It can be stated in two measures -
a. Line pairs per unit distance &
b. Dots(pixels) per unit distance

a. Line pair per unit distance-

• Suppose we construct a chart with alternating black and white vertical lines, each of width
W units (W can be less than 1).
• The width of a line pair is thus 2W and there are 1/2W line pairs per unit distance.
• A widely used definition of image resolution is the largest number of discernible line pairs
per unit distance (e.g. - 100-line pairs per mm)
b. Dots per unit distance is the Measure of image resolution used commonly in printing and
publishing industry and it is expressed as dots per inch (dpi).
Illustrating spatial resolution –

Figure-a, 1250dpi Figure-b, 300dpi

Figure-c, 150dpi Figure-d, 72dpi (checkerboard effect)

The images in Figs. 2.20(a) through (d) are shown in 1250, 300, 150, and 72 dpi, respectively. In order
to facilitate comparisons, all the smaller images were zoomed back to the original size. This is
somewhat equivalent to “getting closer” to the smaller images so that we can make comparable
statements about visible details. There are some small visual differences between Fig(a) and (b), the
most notable being a slight distortion in the large black needle. For the most part, however, Fig.(b) is
quite acceptable. In fact, 300 dpi is the typical minimum image spatial resolution used for book
publishing, so one would not expect to see much difference here. Figure(c) begins to show visible
degradation (see, for example, the round edges of the chronometer and the small needle pointing to 60
on the right side). Figure(d) shows degradation that is visible in most features of the image called
checker board effect.
INTENSITY RESOLUTION à due to quantization
Intensity resolution is a measure of smallest discernible change in intensity level.

Figure-a Figure-b

Figure-c Figure-d

Figure-e Figure-f

Figure-g figure-h
Here, we keep the number of samples constant and reduce the number of intensity levels from 256 to
2, in integer powers of 2. Figure(a) is a 452 * 374 CT projection image, displayed with k= 8 (256
intensity levels). Figures(b) through (h) were obtained by reducing the number of bits k= 7 to k= 1
while keeping the image size constant at 452 * 374 pixels. The 256-, 128-, and 64-level images are
visually identical for all practical purposes. The 32-level image in Fig. (d), however, has an
imperceptible set of very fine ridge-like structures in areas of constant or nearly constant intensity
(particularly in the skull). This effect, caused by the use of an insufficient number of intensity levels
in smooth areas of a digital image, is called false contouring. False contouring generally is quite
visible in images displayed using 16 or less uniformly spaced intensity levels, as the images in Figs.
(e) through (h) show.

False contouring effect - use of insufficient number of intensity levels in smooth areas of a digital
image
Checkerboard effect - because of under sampling of image (insufficient number of pixels)

Image Interpolation
Its a basic tool used in image processing tasks like zooming, shrinking, rotating, geometric corrections.
Whereas Zooming and shrinking task in image are called image resampling.
Interpolation is the process of using known data to estimate values at unknown location.
Three types of interpolation
• Nearest Neighbour interpolation
• Bilinear interpolation
• Bicubic interpolation

Nearest Neighbour interpolation –

It assigns to each new location the intensity of its nearest neighbour in the original image. This
approach is simple but, it has the tendency to produce undesirable artifacts, such as severe distortion
of straight edges.

Bilinear interpolation –
Here, we use the four nearest neighbours to estimate the intensity at a given location. Let (x, y) denote
the coordinates of the location to which we want to assign an intensity value (think of it as a point of
the grid described previously), and let v(x, y) denote that intensity value. For bilinear interpolation,
the assigned value is obtained using the equation

where the four coefficients are determined from the four equations in four unknowns that can be written
using the four nearest neighbours of point (x, y). Bilinear interpolation gives much better results than
nearest neighbour interpolation, with a modest increase in computational burden.

Bicubic interpolation –
It involves the sixteen nearest neighbours of a point. The intensity value assigned to point
(x, y) is obtained using the equation
where the sixteen coefficients are determined from the sixteen equations in sixteen unknowns that can
be written using the sixteen nearest neighbours of point (x, y). Generally, bicubic interpolation does a
better job of preserving fine detail than its bilinear counterpart. Bicubic interpolation is the standard
used in commercial image editing programs, such as Adobe Photoshop and Corel Photopaint.

Some Basic Relationships between Pixels

A pixel p at coordinates (x, y) has four horizontal and vertical neighbours whose coordinates are
given by

This set of pixels, called the 4-neighbors of p, is denoted by N4(p). Each pixel is a unit distance from
(x, y), and some of the neighbour locations of p lie outside the digital image if (x, y) is on the border
of the image.

The four diagonal neighbours of p have coordinates given by

.
and are denoted by ND(p). These points, together with the 4-neighbors, are called the 8-neighbors of
p, denoted by N8(p). As before, some of the neighbour locations in ND(p) and N8(p) fall outside the
image if (x, y) is on the border of the image.

Adjacency, Connectivity, Regions, and Boundaries

Adjacency-
Two pixels that are neighbours and have same grey levels are called adjacent.
Types -
a. 4-adjacency -
Two pixels p and q with values from V are 4-adjacent if q is in the set N4(p).

b. 8-adjacency -
Two pixels p and q with values from V are 8-adjacent if q is in the set N8(p).

c. m-adjacency (mixed adjacency) -

It is introduced to eliminate the ambiguities that often arise when 8-adjacency is used.
Two pixels p and q with values from V are m-adjacent if
(i) q is in N4(p), or
(ii) q is in ND(p) and the set N4(p) ÇN4(q) has no pixels whose value are from V.

Connectivity
Let S represent a subset of pixels in an image. Two pixels p and q are said to be connected in S if
there exists a path between them consisting entirely of pixels in S as shown below.
Regions
Let R be a subset of pixels in an image, we call R a region of the image if R is a connected set as
shown below.

Two regions, Ri and Rj are said to be adjacent if their union forms a connected set as shown below.
Regions that are not adjacent are said to be disjoint.

Boundaries
The boundary of a region R is the set of points that are adjacent to points is the complement of R as
shown below.

Distance Measures
For pixels p, q, and z, with coordinates (x, y), (s, t), and (v, w), respectively, D is a distance function
or metric if

The Euclidean distance between p and q is defined as

The D4 distance (called the city-block distance) between p and q is defined as

The D8 distance (called the chessboard distance) between p and q is defined as