1. Explain the Quantization process with suitable example.

Quantization
Quantization is the process of mapping a range of continuous values to a finite set of discrete
levels. Since digital systems operate with a limited number of values, quantization is necessary
to approximate the sampled data by rounding it to the nearest available level. This process is
a fundamental step in digital image processing, enabling the conversion of an analog image
into a digital format that can be efficiently stored, processed, and transmitted.
Quantization essentially reduces the precision of intensity values, replacing an infinite range
of possible values with a fixed set of predefined levels. However, this approximation
introduces a degree of inaccuracy, known as quantization error, which can cause slight
distortions in the original signal. The extent of this error depends on the number of
quantization levels used—the higher the levels, the more precise the representation of the
original signal.

Key Concepts in Quantization

1. Quantization Levels
Quantization levels refer to the finite number of discrete amplitude values available to
represent the continuous range of intensities. The number of levels is typically determined by
the bit depth of the digital representation.

Impact of Quantization Levels:

o Higher quantization levels result in finer intensity variations, producing a more
accurate representation of the original signal but increasing memory and
processing requirements.
o Lower quantization levels reduce storage needs but lead to a loss of detail,
producing visible artifacts known as banding effects (where smooth transitions
appear as distinct steps).
2. Quantization Error (Quantization Noise)
Quantization error arises due to the difference between the actual continuous intensity values
and their nearest quantized levels. This error is introduced during the rounding process and
results in slight distortions in the processed image.

Effects of Quantization Error:

 When quantization levels are insufficient, fine details and subtle intensity variations
are lost.
 The difference between actual and quantized values appears as noise in an image,
known as quantization noise.
 In images with low quantization levels, uniform color regions might display unnatural
contouring (posterization effect).
Techniques such as dithering (adding random noise before quantization) are sometimes used
to minimize noticeable artifacts in an image.

3. Types of Quantization

Quantization can be categorized based on how the step size (difference between successive
levels) is defined.
A. Uniform Quantization
 The step size remains constant across all intensity values.
 Each quantized level is equally spaced, making it simple to implement.
 Commonly used in digital imaging and video compression.
B. Non-Uniform Quantization
 Step sizes vary across intensity levels.
 More levels are assigned to frequently occurring intensities, reducing quantization
error where needed.
 Often used in applications like audio signal processing and medical imaging.

A common non-uniform quantization technique is logarithmic quantization, which assigns

finer resolution to lower-intensity values while coarsely approximating higher-intensity values.

2. Explain Fundamental Steps in Digital Image Processing

Fundamental Steps in Digital Image Processing
Digital Image Processing (DIP) is a systematic approach to manipulating digital images through
computational methods. It involves multiple stages, each addressing a specific aspect of image
formation, enhancement, transformation, and analysis. These steps collectively allow for
better image representation, extraction of meaningful information, and mathematical
processing for various analytical purposes.

1. Image Acquisition
The image acquisition phase involves capturing an image and converting it into a digital
format. It consists of:
  Sensing: A physical sensor detects electromagnetic radiation reflected or emitted by
an object. The sensor could be an optical device (camera), an infrared detector, or an
X-ray scanner.
 Digitization: The continuous image signal obtained from the sensor undergoes
sampling and quantization to transform it into a discrete numerical representation.
o Sampling involves dividing the image into a grid of discrete pixels.
o Quantization assigns a finite set of intensity values to each pixel, ensuring
digital storage and processing.
The resulting image is represented as a matrix of numerical values corresponding to intensity
levels.

2. Image Preprocessing
Preprocessing is an essential step to improve the quality of an image by suppressing undesired
distortions and enhancing key features.
A. Image Filtering
Filtering is a technique used to modify or enhance certain aspects of an image by processing
pixel values within a defined neighborhood.
 Low-Pass Filtering: Removes high-frequency noise, resulting in a smoothened image.
 High-Pass Filtering: Enhances high-frequency components, such as edges and sharp
transitions.
 Band-Pass Filtering: Retains a specific frequency range while eliminating unwanted
components.
B. Noise Reduction

Various noise filtering techniques exist to mitigate distortions caused by environmental

interference, sensor limitations, or compression artifacts. Common techniques include:
 Linear Filtering (Mean Filter, Gaussian Filter) – Reduces random variations by averaging
neighboring pixel values.
 Non-Linear Filtering (Median Filter, Bilateral Filter) – Preserves edges while eliminating
localized noise.
C. Contrast Enhancement
Contrast manipulation alters intensity values to improve visual distinction between different
image regions. Methods include:
 Histogram Equalization: Spreads pixel intensities to enhance overall contrast.
  Gamma Correction: Adjusts brightness distribution non-linearly to emphasize specific
intensity ranges.

3. Image Restoration
Image restoration aims to reconstruct or recover an image that has been degraded by known
distortions. Unlike enhancement, restoration relies on mathematical models of the
degradation process.

B. Restoration Techniques
 Inverse Filtering: Attempts to reverse the degradation by applying the inverse of H.
 Wiener Filtering: Uses statistical methods to minimize mean square error between the
restored and original image.
 Blind Deconvolution: Estimates both the degradation function and the original image
when H is unknown.

4. Color Image Processing

Color image processing extends monochrome image techniques by incorporating multiple

color channels. This involves processing images in different color spaces to achieve desired
transformations.
A. Color Models

 RGB (Red, Green, Blue): Represents images as a combination of three primary colors.
 CMY (Cyan, Magenta, Yellow): Used in subtractive color processing.
 HSI (Hue, Saturation, Intensity): Separates chromatic content from brightness, allowing
for effective color segmentation.
B. Color Transformation
Operations on color images include:
 Color Normalization: Standardizes intensity across different lighting conditions.

 Color Filtering: Isolates specific hues for further analysis.

5. Wavelet Transform in Image Processing

Wavelet transform is a multiresolution analysis technique that decomposes an image into
different frequency components while preserving spatial information. Unlike Fourier
Transform, which provides only frequency domain representation, wavelet transform captures
both time-domain (spatial) and frequency-domain characteristics.
A. Mathematical Representation
Wavelet decomposition involves convolving an image with basis functions called wavelets. The
transformed image contains multiple levels of detail coefficients, representing different
frequency bands.
B. Types of Wavelet Transforms
 Continuous Wavelet Transform (CWT): Provides a continuous representation over all
scales.
 Discrete Wavelet Transform (DWT): Decomposes images into approximation and detail
coefficients using filters.
 Haar Wavelet Transform: One of the simplest wavelets, commonly used in image
compression.

6. Image Compression
Image compression reduces the amount of data required to store or transmit an image by
eliminating redundant or non-essential information.
A. Compression Categories
 Lossless Compression: Preserves all original information without degradation.
Common algorithms include Huffman coding and Run-Length Encoding (RLE).
 Lossy Compression: Reduces data size by discarding perceptually less significant
information. The Discrete Cosine Transform (DCT) used in JPEG is a prime example.
B. Compression Techniques
 Transform Coding: Converts image data into frequency components using DCT or
wavelet transform.
 Predictive Coding: Estimates pixel values based on neighboring pixels.

7. Morphological Image Processing

Morphological operations analyze and process image structures based on shape and
connectivity. These operations are primarily applied to binary images.
A. Fundamental Morphological Operations
 Erosion: Removes pixels from object boundaries.
 Dilation: Expands object boundaries by adding pixels.

 Opening & Closing: Used to remove noise while preserving the main structure of the
image.

8. Image Segmentation
Segmentation partitions an image into distinct regions based on pixel attributes such as
intensity, color, and texture.

9. Feature Extraction
Feature extraction identifies key characteristics of an image that can be used for classification
or recognition.
A. Types of Features
 Statistical Features: Measure properties such as mean, variance, and entropy.
  Geometrical Features: Extract shape descriptors such as edges, contours, and corners.
 Texture Features: Capture patterns using methods like Gabor filters and Gray-Level Co-
Occurrence Matrices (GLCM).

10. Image Pattern Recognition

Pattern recognition involves classifying image regions based on extracted features.

Elements of Visual Perception

3. What are the components of digital image processing system? Explain each in detail.

Components of a Digital Image Processing System

A Digital Image Processing (DIP) System is composed of various interconnected components that
work together to acquire, process, and analyze digital images. Each component plays a critical role in
transforming raw image data into a processed form suitable for further interpretation or storage. The
components of a DIP system include:

1. Image Acquisition Device (Sensors and Digitizer)

2. Storage Unit

3. Processing Unit

4. Software and Algorithms

5. Display Device

6. Communication System

Each of these components is essential for effective image processing and contributes to different
aspects of image handling. Below is a detailed explanation of each component:
1. Image Acquisition Device (Sensors and Digitizer)

A. Image Sensors

The first step in digital image processing involves capturing an image through an image sensor. An
image sensor is a device that detects and collects light or electromagnetic radiation from the scene
and converts it into an electrical signal.

Types of image sensors include:

 Charge-Coupled Device (CCD) Sensors: These sensors contain an array of photodetectors

that convert light into electrical signals. CCDs are known for their high sensitivity and low
noise levels.

 Complementary Metal-Oxide-Semiconductor (CMOS) Sensors: These sensors integrate

photodetectors and amplification circuits on a single chip, offering lower power consumption
and faster processing speeds compared to CCDs.

 Infrared (IR) Sensors: Used for thermal imaging by detecting heat radiations.

 X-ray Sensors: Capture high-energy radiations for medical and industrial applications.

B. Digitizer (Analog-to-Digital Conversion)

Once the sensor captures the image as an analog signal, it must be digitized for computational
processing. The digitization process consists of two main steps:

1. Sampling: Converts the continuous image into a discrete grid of pixels. The resolution of the
image depends on the number of sampled points.

2. Quantization: Assigns a finite number of intensity levels to each sampled pixel. Higher
quantization levels result in better image representation.

The digitizer converts the continuous analog signal into a discrete numerical format, making it
suitable for digital computation.

2. Storage Unit

After acquisition, the digital image needs to be stored in a suitable medium for further processing or
retrieval. The storage unit manages the efficient handling of image data.

A. Primary Storage (Volatile Memory - RAM)

 Temporary storage for images being processed.

 Ensures quick access to pixel data for real-time applications.

B. Secondary Storage (Non-Volatile Memory - Hard Disks, SSDs, Cloud Storage)

 Stores processed and unprocessed images for long-term use.

 Data compression techniques such as JPEG (lossy) and PNG (lossless) are often applied to
optimize storage.

C. Buffer Memory
  Holds temporary image data before being transferred between processing and storage units.

Storage devices ensure efficient data retrieval, modification, and archiving for further processing.

3. Processing Unit

The processing unit is the core computational component responsible for executing image processing
operations. It consists of:

A. Central Processing Unit (CPU)

 Performs arithmetic and logical operations on image data.

 Handles basic image manipulations such as brightness adjustments, filtering, and

transformations.

B. Graphics Processing Unit (GPU)

 Specialized hardware for parallel processing of images.

 Used for high-speed operations such as convolution, edge detection, and neural network
computations.

C. Digital Signal Processor (DSP)

 Optimized for performing mathematical computations on digital signals.

 Used in real-time applications where speed and efficiency are critical.

The processing unit executes algorithms such as image enhancement, restoration, segmentation, and
pattern recognition.

4. Software and Algorithms

Software plays a vital role in defining the methods and techniques used for processing images. Image
processing software consists of various algorithms designed to manipulate pixel data mathematically.

A. Low-Level Processing Algorithms

 Filtering Algorithms: Used for noise reduction and edge enhancement (e.g., Gaussian Filter,
Median Filter).

 Histogram Equalization: Adjusts contrast by redistributing intensity levels.

 Image Transformation: Includes Fourier Transform and Wavelet Transform for frequency
analysis.

B. Mid-Level Processing Algorithms

 Segmentation Techniques: Partition an image into meaningful regions (e.g., Watershed

Algorithm, Region Growing).

 Feature Extraction: Identifies significant image characteristics such as edges, textures, and
shapes.
C. High-Level Processing Algorithms

 Pattern Recognition and Classification: Used in facial recognition, medical diagnostics, and
automated inspection.

 Artificial Intelligence (AI) and Machine Learning Models: Neural networks, convolutional
neural networks (CNNs), and deep learning architectures improve object detection and scene
interpretation.

Software packages such as MATLAB, OpenCV (Python & C++), and TensorFlow are commonly used for
implementing image processing techniques.

5. Display Device

The processed image must be visualized on an appropriate display device for interpretation. The
quality of the display significantly impacts the accuracy of image analysis.

A. Monitors and Screens

 Cathode Ray Tube (CRT) Displays: Older technology with low resolution and high power
consumption.

 Liquid Crystal Display (LCD) and Light Emitting Diode (LED) Screens: Modern displays
offering high resolution, better contrast, and low power consumption.

 Organic LED (OLED) and AMOLED Displays: Provide deeper blacks and higher contrast ratios
for improved visualization.

B. Projectors

 Used in large-scale visualization applications such as medical imaging, remote sensing, and
scientific research.

C. Specialized Display Systems

 Medical Monitors: High-resolution screens used in radiology for accurate analysis of medical
scans (e.g., X-rays, MRI).

 Virtual Reality (VR) and Augmented Reality (AR) Displays: Enable interactive image
exploration.

The display system plays a crucial role in human interpretation and decision-making based on
processed images.

6. Communication System

Image processing systems often involve data transfer across networks for sharing, analysis, and
storage. The communication system facilitates the transmission of image data between different
components.

A. Wired Communication

 Ethernet and Optical Fiber: High-speed data transfer for industrial and medical imaging.
  USB and HDMI: Used for connecting image acquisition devices to processing units.

B. Wireless Communication

 Wi-Fi and Bluetooth: Enable remote access and cloud-based image storage.

 5G and IoT Connectivity: Enhances real-time transmission of high-resolution images in

telemedicine and surveillance.

4. Explain in detail about image acquisition system.

Image Acquisition

Image acquisition is the first and most fundamental step in digital image processing. It refers to the
process of capturing an image using a sensor, converting it into digital form, and storing it for further
processing. The quality and accuracy of digital images heavily depend on the method of acquisition,
as it determines the level of detail, resolution, and color fidelity.

Image acquisition systems rely on the electromagnetic spectrum, which includes different forms of
radiation such as visible light, infrared, X-rays, and gamma rays. The selection of an acquisition
method depends on the specific application, whether it is for medical imaging, satellite surveillance,
industrial inspection, or consumer photography.

1. Electromagnetic Spectrum and Image Acquisition

The electromagnetic spectrum plays a critical role in image acquisition as different imaging
technologies use different ranges of wavelengths to capture images. The commonly used regions of
the electromagnetic spectrum in imaging include:

 Visible Light (400-700 nm): Used in photography, webcams, and digital cameras.

 Infrared (700 nm - 1 mm): Used in night vision cameras and thermal imaging.

 Ultraviolet (10 nm - 400 nm): Used in biological imaging and forensics.

 X-rays (0.01 nm - 10 nm): Used in medical imaging such as radiography and CT scans.

 Gamma Rays (less than 0.01 nm): Used in nuclear medicine and astrophysical imaging.
Each of these imaging techniques requires specialized sensors that are sensitive to the corresponding
wavelength range.

2. Image Sensors for Digital Image Acquisition

An image sensor is a device that captures light and converts it into an electrical signal that can be
processed digitally. The two most common types of sensors used in digital imaging are:

A. CMOS (Complementary Metal Oxide Semiconductor) Sensors

CMOS sensors are widely used in digital cameras, smartphones, and webcams due to their
advantages in speed, cost, and energy efficiency.
Key Features:

 Each pixel in a CMOS sensor has its own analog-to-digital converter (ADC), allowing for faster
processing.

 CMOS sensors consume less power, making them ideal for battery-powered devices.

 They offer higher frame rates, making them suitable for video recording.

 However, they can suffer from rolling shutter effects and may have lower sensitivity
compared to CCD sensors.

B. CCD (Charge-Coupled Device) Sensors

CCD sensors were once the dominant technology in digital cameras and high-end imaging
applications.

Key Features:

 CCD sensors use a single ADC, shifting pixel charges across the sensor before converting
them to digital values.

 They provide superior image quality with lower noise and better light sensitivity.

 However, they consume more power and are generally slower than CMOS sensors.

 Due to advancements in CMOS technology, CCD sensors are now less commonly used.

3. Color Image Acquisition

Human vision perceives colors based on three primary wavelengths—red, green, and blue (RGB).
Digital imaging systems use color filters to separate light into these primary components. The most
commonly used method for color acquisition is the Bayer filter, which consists of a mosaic pattern of
red, green, and blue filters.

Key Features of the Bayer Filter:

 The filter is arranged in a 2x2 matrix, with two green, one red, and one blue filter per unit.

 Green is more dominant because human eyes are more sensitive to green wavelengths.

 A demosaicing algorithm is used to reconstruct full-color images from the captured sensor
data.

Other color acquisition methods include:

 Three-chip cameras: Use separate CCD or CMOS sensors for red, green, and blue channels,
providing higher accuracy but at a higher cost.

 Foveon X3 sensors: Capture full-color information at each pixel location without requiring a
Bayer filter.

4. Sampling in Image Acquisition

Once an image is captured, it must be converted from its continuous analog form into a discrete
digital format. This process is known as sampling.

Definition of Sampling

Sampling refers to the process of selecting specific points from a continuous signal to create a
discrete representation. In digital imaging, this means dividing an image into small units called pixels,
each representing a specific intensity value.

Key Concepts in Sampling:

1. Sampling Rate (Resolution): The number of pixels per unit area in an image. Higher sampling
rates result in better image resolution.

2. Aliasing: A distortion that occurs when the sampling rate is too low, causing fine details to
appear incorrectly. This is prevented using an anti-aliasing filter before sampling.

3. Types of Sampling:

o Uniform Sampling: Pixels are evenly spaced at fixed intervals.

o Non-Uniform Sampling: Pixels are placed at varying intervals, often used in

specialized applications like medical imaging.

o Flat-Top Sampling: Pixels have constant amplitude over a small duration.

5. Quantization in Image Acquisition

After sampling, the intensity values of an image must be mapped to a limited set of discrete levels, a
process known as quantization.

Definition of Quantization

Quantization involves rounding continuous intensity values to the nearest discrete level, allowing
images to be represented in digital form. The accuracy of quantization depends on the bit depth of
the image.

Key Concepts in Quantization:

1. Quantization Levels: The number of distinct intensity values available, determined by the bit
depth.

o 8-bit images have 256 levels (0-255).

o 16-bit images have 65,536 levels.

2. Quantization Error: The difference between the actual and quantized values, which
introduces distortion in the image.

3. Types of Quantization:

o Uniform Quantization: Equal step sizes between levels, simple but can introduce
artifacts.

o Non-Uniform Quantization: Variable step sizes, used for more efficient encoding in
applications like audio and medical imaging.
Higher bit-depth images reduce quantization error but require more storage and processing power.

6. Factors Affecting Image Acquisition Quality

Several factors influence the quality of the acquired image:

 Lighting Conditions: Poor lighting can lead to noise and lower contrast.

 Sensor Quality: Higher-quality sensors provide better dynamic range and sensitivity.

 Lens and Optics: Optical distortions can degrade image clarity.

 Motion Artifacts: Fast movement can cause blurring or rolling shutter effects.

 Environmental Interference: External electromagnetic signals can introduce noise into the
image.

5. Explain different logarithmic transformations with examples.

Logarithmic Transformations

Logarithmic transformations are a class of intensity transformations that apply a logarithmic function
to image pixel values. These transformations are useful for enhancing details in images with a wide
range of intensity values, such as medical images, astronomical images, and low-light photographs.

The general form of a logarithmic transformation is:

s=clog⁡(1+r)s = c \log(1 + r)s=clog(1+r)

Where:

 sss is the output pixel intensity,

 rrr is the input pixel intensity,

 ccc is a scaling constant to normalize the output, and

 log⁡(1+r)\log(1 + r)log(1+r) ensures that the logarithm is always defined (as log(0) is
undefined).

Logarithmic transformations help to expand dark intensity values while compressing bright intensity
values, making them particularly useful for images with high contrast.

1. Types of Logarithmic Transformations

A. Log Transformation

The log transformation is defined by the formula:

s=clog⁡(1+r)s = c \log(1 + r)s=clog(1+r)

This transformation enhances low-intensity pixels more than high-intensity pixels. It is useful for
revealing details in dark areas of an image while compressing the brighter areas.
Example

Consider an image where a large portion is underexposed (dark regions). Applying a log
transformation will brighten the dark regions while keeping the already bright regions from
becoming too intense.

Effect on Image

 Dark pixels (low values of rrr) are significantly brightened.

 Bright pixels (high values of rrr) are slightly compressed to avoid saturation.

 Used in medical imaging (X-rays) and satellite imagery to enhance dark details.

B. Inverse Log (Anti-Log) Transformation

The inverse log transformation is defined by:

s=c(10r−1)s = c (10^r - 1)s=c(10r−1)

This transformation expands bright intensity values while compressing dark values. It is used when
we need to emphasize bright areas in an image.

Example

Suppose we have an image with very bright regions (like a sunlit sky) and some details lost in
highlights. The inverse log transformation helps to enhance these bright areas, making them more
distinguishable.

Effect on Image

 Bright pixels (high values of rrr) are stretched, making them even brighter.

 Dark pixels (low values of rrr) remain relatively unchanged.

 Useful in medical imaging, where bright regions (such as bones in X-rays) need
enhancement.

Applications

 Medical Imaging: Enhancing details in low-contrast X-ray or MRI scans.

 Satellite Imagery: Revealing details in dark terrains while keeping bright clouds controlled.

 Astrophotography: Bringing out faint celestial objects in space images.

 Forensics and Security: Enhancing low-light surveillance footage.

6. Explain the histogram equalization method of image enhancement with example.
Histogram Equalization

Histogram Equalization is a widely used image enhancement technique that improves the contrast of
an image by redistributing the intensity values. It is particularly useful for images that are too dark
or too bright, where details are not easily visible due to poor contrast.

The primary goal of histogram equalization is to spread out the intensity levels in an image so that it
makes full use of the available dynamic range. This leads to an image with better contrast and more
visible details.

1. Understanding Histograms in Images

A histogram represents the distribution of intensity values in an image.

 A dark image has most pixel intensities concentrated in the lower (left) part of the
histogram.

 A bright image has most pixel intensities concentrated in the higher (right) part of the
histogram.

 A low-contrast image has intensity values clustered around a narrow range.

Histogram Equalization spreads out these intensity values to utilize the full range [0,255][0,
255][0,255] in an 8-bit image, making the image more balanced and visually appealing.

2. Steps in Histogram Equalization

Step 1: Compute the Histogram

 Count the number of pixels for each intensity level in the image.

 This gives the frequency distribution of intensity values.

Step 2: Compute the Probability Density Function (PDF)

 The PDF is calculated as:

P(rk)=Number of pixels with intensity rkTotal number of pixelsP(r_k) = \frac{\text{Number of pixels

with intensity } r_k}{\text{Total number of pixels}}P(rk
)=Total number of pixelsNumber of pixels with intensity rk

where rkr_krk is an intensity level.

Step 3: Compute the Cumulative Distribution Function (CDF)

 The CDF is calculated as:

CDF(rk)=∑j=0kP(rj)CDF(r_k) = \sum_{j=0}^{k} P(r_j)CDF(rk)=j=0∑kP(rj)

 The CDF accumulates the probability values and helps in mapping old pixel values to new
ones.

Step 4: Compute the New Intensity Levels

  The new intensity values sks_ksk are obtained using:

sk=(max intensity)×CDF(rk)s_k = (\text{max intensity}) \times CDF(r_k)sk=(max intensity)×CDF(rk)

o For an 8-bit image, max intensity is 255.

o The result is rounded to the nearest integer.

Step 5: Map the Old Intensities to New Ones

 Each pixel in the original image is replaced with its new intensity value sks_ksk from the
computed mapping.

3. Example of Histogram Equalization

Original Image Histogram

Consider an image with pixel intensity values mostly between 50 and 150. This means the histogram
is concentrated in this range, leading to poor contrast.

Intensity Level Frequency (Number of Pixels)

0-50 Low

50-150 High (Clustered)

150-255 Low

After Histogram Equalization

The intensity values are redistributed across the full range (0 to 255), leading to improved contrast.

Intensity Level Frequency (Number of Pixels)

0-255 Evenly Distributed

This means dark areas become lighter and bright areas become darker, creating better overall
contrast.

4. Applications

1. Medical Imaging – Enhances details in X-rays and MRI scans.

2. Satellite Imaging – Improves visibility of land and water features.

3. Forensic Image Processing – Enhances dark or low-light surveillance images.

4. Photography & Image Processing – Used to enhance photos with poor contrast.
 7. Explain about affine transformation in image processing.

Affine Transformation

Affine transformation is a fundamental geometric transformation used in image processing that

preserves collinearity (points on a line remain on the same line) and ratios of distances (midpoints
remain midpoints). It allows for scaling, rotation, translation, shearing, and reflection of an image
while maintaining the parallelism of lines.

Affine transformation is widely used in image registration, object detection, computer vision, and
graphics processing.
3. Properties of Affine Transformation

1. Preserves collinearity – Straight lines remain straight.

2. Preserves parallelism – Parallel lines remain parallel.

3. Does not preserve distances and angles – Shapes may be distorted, but relative proportions
are maintained.

4. Compositions of affine transformations are also affine – Combining multiple

transformations still results in an affine transformation.

5. Applications

1. Image Rotation and Scaling – Used in image editors and graphic design software.

2. Object Detection & Recognition – Helps in aligning objects for feature extraction.

3. Face Recognition – Used to normalize face images for better matching.

4. Augmented Reality (AR) – Helps in overlaying virtual objects in the real world.

5. Geometric Corrections in Satellite Images – Used to correct distortions due to sensor

movement.
8. Compute the Euclidean Distance (D1), City-block Distance (D2) and Chessboard
distance (D3) for points p and q, where p and q be (5, 2) and (1, 5) respectively. Give
answer in the form (D1, D2, D3).