Discuss the processes of quadtree splitting and merging in image segmentation (3
marks)
Splitting:
● The image is repeatedly divided into four equal-sized square regions (quadrants).
● This process continues until each region becomes homogeneous (i.e., pixels within the
region have similar intensity or color).
● It helps identify uniform areas in the image.
Merging:
● Adjacent regions that have similar properties are combined (merged) to form larger
homogeneous regions.
● This reduces over-segmentation and produces more meaningful segments.
Result:
● The combination of splitting and merging results in a hierarchical representation of the
image that efficiently segments it into uniform regions.
Illustrate and describe the calculation of parameters, in a triangular plane , of the colour
model which closely resembles the human perception
Colour Model Closely Resembling Human Perception – CIE Chromaticity
Diagram (CIE 1931)
1. Illustration (Triangular Plane)
● The CIE chromaticity diagram is a triangular plane where each point represents a
specific colour.
● The triangle’s vertices correspond to the primary colours (Red, Green, Blue).
● Any colour inside the triangle can be represented as a mixture of these three
primaries.
● The diagram shows how the human eye perceives colour, based on the response of
the three types of cone cells.
2. Calculation of Parameters
The colour of a light source is described by its tristimulus values (X, Y, Z) obtained using
colour-matching functions.
From these, the chromaticity coordinates are calculated as:
3. Description
● x, y coordinates specify the hue and saturation (chromaticity).
● Y represents luminance (brightness).
● This model accurately represents human visual perception of colour differences.
Discuss the two pseudo colour image processing techniques
1. Intensity Slicing (Gray Level Slicing)
● Concept:
The range of gray levels (intensity values) in an image is divided into several intervals
(slices).
Each interval is then assigned a specific colour.
● Purpose:
It helps in emphasizing certain ranges of intensities—for example, different temperature
levels in a thermal image or elevation levels in a topographic map.
● Process:
○ Divide the gray level range (0–255) into a number of slices.
○ Assign a colour to each slice.
○ Display the resulting coloured image.
● Example:
○ Pixels with intensity 0–50 → Blue
○ 51–150 → Green
○ 151–255 → Red
2. Colour Transformation (Colour Mapping)
● Concept:
A transformation function maps each gray level to a specific colour according to a
chosen colour model (e.g., RGB or HSV).
● Purpose:
To apply a smooth transition of colours across intensity levels, giving a visually
continuous representation (unlike discrete bands in intensity slicing).
● Process:
○ Define mapping functions R(f),G(f),B(f)R(f), G(f), B(f)R(f),G(f),B(f) for each
intensity value fff.
○ Combine them to generate the pseudo colour image.
● Example:
○ A “rainbow” colour map where low intensities appear blue, mid-range intensities
appear green/yellow, and high intensities appear red.
What are descriptors? How can simple descriptors be used for boundary description
purpose ? (3 marks)
Descriptors are quantitative measurements or features used to describe the shape, size, and
other characteristics of objects in an image.
They help in object recognition, classification, and analysis in image processing.
Use of Simple Descriptors for Boundary Description:
Boundary descriptors describe the shape of an object based on its boundary or contour
rather than its entire region.
Simple descriptors used for boundary description include:
1. Perimeter:
○ The total length of the object’s boundary.
○ Calculated by counting the number of boundary pixels or using Euclidean
distance between successive boundary points.
2. Compactness (Form factor):
○ Measures how closely the shape resembles a circle.
○ Formula:
C = P^2/4*pi*A
○ where P = perimeter, A = area.
○ A perfect circle has the smallest compactness value.
3. Eccentricity / Aspect Ratio:
○ Ratio of major axis to minor axis of the object’s boundary.
○ Indicates how elongated the shape is.
Define hue, saturation and chromaticity (3 marks)
1. Hue:
○ Refers to the dominant wavelength or pure colour perceived by the human eye
(e.g., red, green, blue, yellow).
○ It indicates the type of colour in the image and is independent of brightness.
2. Saturation:
Represents the purity or vividness of a colour.
○ A high saturation colour appears rich and pure, while a low saturation colour
appears dull or grayish.
○ It defines how much the colour is diluted with white light.
3. Chromaticity:
○ Describes the quality of colour irrespective of its brightness.
○ It is defined by hue and saturation together, representing the colour’s tone and
purity but not its intensity.
Short note on 1. region growing 2. region splitting and merging (6 mark)
1. Region Growing
● Definition:
Region growing is a pixel-based image segmentation technique that groups pixels or
subregions into larger regions based on similarity criteria such as intensity, colour, or
texture.
● Process:
○ Select one or more seed points within the image.
○ Compare neighbouring pixels to the seed based on similarity (e.g., gray level
difference).
If the pixel is similar, add it to the region.
○ Continue the process until no more pixels satisfy the condition.
● Advantages:
○ Produces connected regions.
○ Simple and effective for homogeneous regions.
● Disadvantages:
○ Sensitive to noise and choice of seed points.
○ Computationally intensive for large images.
● Applications:
Used in medical image segmentation, object detection, and texture analysis.
2. Region Splitting and Merging
● Definition:
A top-down and bottom-up segmentation technique that divides an image into uniform
regions and then merges adjacent regions that meet a homogeneity criterion.
● Process:
○ Start with the entire image as a single region.
○ Split the region into four quadrants (quadtree) if it is not homogeneous.
○ Continue splitting until all subregions are homogeneous.
○ Merge adjacent homogeneous regions to form larger regions.
● Advantages:
○ No need for seed points.
○ Reduces over-segmentation by merging similar areas.
● Disadvantages:
○ Computationally more complex.
○ Boundary accuracy may be lower than pixel-based methods.
● Applications:
Used in remote sensing, object-based image analysis, and texture segmentation.
What is a structuring element ? How is it used for dilation operation? (3 marks)
A structuring element is a small matrix or shape used in morphological image processing
to probe or interact with an image.
It defines the neighborhood over which an operation (like dilation or erosion) is performed.
● It has:
○ A specific shape (e.g., square, circle, cross).
○ A defined origin (reference point), usually at its center.
Use in Dilation Operation:
● Dilation adds pixels to the boundaries of objects in a binary image.
● For each pixel in the image, the structuring element is placed such that its origin
coincides with that pixel.
● If any pixel of the structuring element overlaps with a foreground (white) pixel in the
image, the corresponding output pixel is set to 1 (white)
Explain the RGB and HSI colour model in brief (8 marks)
1. RGB Colour Model (Red, Green, Blue)
● Definition:
The RGB model is an additive colour model in which colours are formed by combining
the three primary colours — Red (R), Green (G), and Blue (B) — in different
proportions.
● Working principle:
○ When R, G, and B = 0, the result is black.
○ When R, G, and B = 1 (maximum), the result is white.
○ Intermediate values create different colours (e.g., R=1, G=0, B=0 → red).
● Applications:
Used in display devices like monitors, TVs, cameras, and scanners where light is
emitted.
● Advantages:
○ Simple to implement.
○ Directly matches how electronic devices generate colours.
2. HSI Colour Model (Hue, Saturation, Intensity)
● Definition:
The HSI model represents colour in a way that matches human visual perception,
separating the chromatic content (hue and saturation) from the brightness (intensity).
● Components:
○ Hue (H): Represents the colour type (e.g., red, green, blue) — measured as an
angle (0°–360°) on a colour wheel.
○ Saturation (S): Indicates purity or vividness of colour — 0 means grey, 1
means pure colour.
○ Intensity (I): Represents brightness — average of RGB components.
● Applications:
Used in image processing, segmentation, and computer vision, where human
perception is important.
● Advantages:
○ Separates colour information from brightness.
○ Easier for colour-based image analysis.
What is a chain code ? explain (4 mark)
Chain Code – Definition & Explanation
● Definition:
A chain code is a method of representing a boundary or contour of a digital image
object by using a sequence of directional codes that describe the path of the
boundary pixels.
Explanation:
● The boundary of an object is traced in a clockwise or counterclockwise direction.
● Each movement from one boundary pixel to the next is assigned a direction code
based on a reference grid.
● There are two common types:
○ 4-connected chain code: directions are 0, 1, 2, 3 → (right, up, left, down)
○ 8-connected chain code: directions are 0–7 representing all eight possible
directions (including diagonals).
Write short note on :
1. texture 2. fourier descriptors 3. Intensity slicing (9 mark)
1. Texture (3 marks)
● Definition:
Texture refers to the visual pattern or surface property of an image region that
describes how pixel intensities are spatially arranged.
It gives information about the structure and appearance of objects, such as
smoothness, roughness, or regularity.
● Types of texture features:
○ Statistical: Based on pixel intensity relationships (e.g., contrast, entropy).
○ Structural: Based on repetition of basic texture elements.
○ Spectral: Based on frequency domain properties.
● Applications:
Used in image segmentation, pattern recognition, surface inspection, and remote
sensing.
2. Fourier Descriptors (3 marks)
● Definition:
Fourier descriptors are shape representation techniques obtained by applying the
Fourier Transform to the boundary coordinates of an object.
● Explanation:
○ The boundary of an object is represented as a sequence of complex numbers.
○ The Discrete Fourier Transform (DFT) is applied to this sequence to obtain
frequency components.
○ These coefficients are called Fourier Descriptors.
○ Lower frequency descriptors capture the general shape, while higher ones
represent fine details.
● Applications:
Useful in shape analysis, object recognition, and image classification since they are
rotation, scale, and translation invariant.
3. Intensity Slicing (3 marks)
● Definition:
Intensity slicing is a technique used to highlight specific intensity ranges in a
grayscale image by assigning different colours or shades to different intensity levels.
● Explanation:
○ The grayscale range (0–255) is divided into several slices (intervals).
○ Each slice is then displayed using a distinct colour or brightness level.
○ This helps to visualize hidden details or features that are not easily visible in
the original image.
● Applications:
○ Medical imaging (e.g., highlighting tissues or tumors)
○ Remote sensing (e.g., land cover classification)
○ Image enhancement
○
○
○
Expllain CMY colour model used in Digital Image Processing
1. Definition:
The CMY colour model is a subtractive colour model used mainly in printing and image
reproduction.
It represents colours as combinations of the three secondary colours —
Cyan (C), Magenta (M), and Yellow (Y) — which are the complements of Red, Green, and
Blue (RGB).
2. Principle:
● In the subtractive model, colours are formed by absorbing (subtracting) light.
● Each component absorbs its complementary primary colour:
○ Cyan absorbs Red
○ Magenta absorbs Green
○ Yellow absorbs Blue
Thus, by combining these three inks or filters, various colours are produced by
subtracting portions of white light.
3. Mathematical Relationship with RGB:
The CMY model is derived from the RGB model as:
C=1-R
M=1-G
Y=1-B
where R, G, B, C, M, Y are normalized between 0 and 1.
4. Working Example:
● White light (R=G=B=1) → C=M=Y=0 (no ink).
● Black (R=G=B=0) → C=M=Y=1 (all inks combined).
5. Applications:
● Colour printing and publishing systems (printers, scanners).
● Digital image processing for converting display (RGB) images to print formats
(CMY or CMYK).
● Colour correction and reproduction tasks.
6. Extension – CMYK Model:
In practice, a fourth component (K) for black is added → CMYK model, because mixing C,
M, and Y doesn’t produce a pure black.
Define Intensity (2 mark)
Intensity – Definition
● Definition:
Intensity refers to the brightness level of a pixel in a digital image.
It represents the amount of light energy received or reflected by that pixel
●
● In a grayscale image, intensity is a single value ranging from 0 (black) to 255
(white).
● In a colour image, intensity is often the average or weighted sum of the red, green,
and blue components.
I=R+G+B/3
Define Shape Number
Shape Number (5 Marks)
1. Definition:
A Shape Number is a numerical descriptor used to represent the shape or boundary of an
object in a digital image.
It is derived from the chain code of the object’s boundary and provides a compact and
rotation-invariant representation of its shape.
2. Concept:
● When tracing the boundary of an object, a chain code is generated (sequence of
direction numbers).
● The Shape Number is obtained by taking the first difference between successive
chain code directions, measured mod N, where N = 4 or 8 (for 4-connected or
8-connected boundaries).
Compute the shape number for a hexagon image segment ?
Define Euler number (3 mark)
Euler Number – Definition (3 Marks)
● Definition:
The Euler Number is a topological property of a binary image that gives a
measure of the connectivity of objects within the image.
It is defined as:
E=C−H
where
C = Number of connected components (objects)
H = Number of holes within those objects
Draw HIS circular colour plane . Mark hue, saturation and intensity of a colour point and
explain
(Then write about hue, saturation and intensity)
Discuss the pseudocolour image processing
1. Gray level transformation
2. Intensity to colour transformation
1. Gray Level (Intensity) Transformation
Concept:
In this method, the gray levels of the image (ranging from 0 to 255) are mapped to
specific colours.
Each intensity value or range of values is assigned a colour according to a predefined
colour mapping function.
Techniques:
● Intensity slicing (Gray level slicing):
The gray scale is divided into slices (ranges), and each range is assigned a
different colour.
Example: Dark areas (low intensity) may appear blue, mid-gray may appear green,
and bright regions may appear red.
● Piecewise Linear Transformation:
Specific gray ranges are mapped linearly to colour values, giving a smooth colour
transition.
Application Example:
Used in medical imaging (like X-rays) or satellite images to highlight features such as
temperature or elevation differences.
2. Intensity-to-Colour Transformation
Concept:
This approach uses a continuous mathematical transformation that converts the
intensity value (I) into a set of colour components (R, G, B) or (H, S, I) using a colour
model.
Methods:
1. Function-based Transformation:
○ Each RGB component is generated as a function of intensity I:
○ R = f1 (I) , G = f2 (I) , B = f3 (I)
○
Example:
■ Low intensity → Blue
■ Medium intensity → Green
■ High intensity → Red
2. Colour Model-based Transformation (HSI or HSV):
Intensity is mapped to Hue (colour type), while Saturation and Brightness
are kept constant.
Example: Dark gray → Blue hue; light gray → Yellow hue.
Applications:
● Weather maps (temperature visualization)
● Thermal imaging
● Elevation maps
● Medical diagnostics
Which colour model uses subtractive mixing ?explain
Colour Model Using Subtractive Mixing – CMY Colour Model (3 Marks)
● The CMY (Cyan, Magenta, Yellow) colour model uses subtractive colour mixing.
● In this model, colours are created by absorbing (subtracting) certain wavelengths
of white light using pigments or inks.
○ Cyan absorbs Red
○ Magenta absorbs Green
○ Yellow absorbs Blue
● It is the inverse of the RGB model, defined as:
C=1−R, M=1−G, Y=1−B
● Application: Used in colour printing and image reproduction systems (printers
and scanners).