SRM University AP, Andhra Pradesh

Neerukonda, Mangalagiri Mandal,

Guntur District, Mangalagiri, Andhra Pradesh – 522240.

Digital Image Processing

L T P C
Course Code CSE 456 Course Category Stream Elective (SE)
3 0 1 4
Pre-Requisite Progressive
CSE 336 Co-Requisite Course(s)
Course(s) Course(s)
Course Offering Professional / Licensing MathWorks License
Department CSE Standards for MATLAB software

Course Objectives / Course Learning Rationales (CLRs)

1. Understand the overview of the field of image processing.
2. Gain knowledge of the fundamental algorithms and how to implement them.
3. Prepare to read the current image processing research literature.
4. Gain experience in applying image processing algorithms to real problems.

Course Outcomes / Course Learning Outcomes (CLOs)

Expected Expected
Bloom’s
At the end of the course the learner will be able to Proficiency Attainment
Level
Percentage Percentage
Describe the process of image processing and techniques
Outcome 1 2 75% 75%
involved in image processing pipeline.
Outcome 2 Identify image enhancement techniques. 2 75% 70%
Illustrate the causes for image degradation and overview of
Outcome 3 3 70% 65%
image restoration techniques.
Apply spatial and frequency domain techniques
Outcome 4 3 70% 65%
for image compression.
Demonstrate extraction techniques for image analysis and
Outcome 5 3 75% 70%
recognition.
Develop an image processing application using feature
Outcome 6 5 65% 60%
extraction and representation
Outcome 7 Recognize the rapid advances in Machine vision. 2 70% 65%

Course Articulation Matrix (CLO) to Program Learning Outcomes (PLO)

Program Learning Outcomes (PLO)

Project Management
Analysis, Design and

Multicultural Skills

Life Long Learning

Moral, and Ethical
Environment and

Self-Directed and
Modern Tool and

Teamwork Skills

Communication
Individual and
Sustainability
Development

and Finance
Engineering

Society and
Knowledge

Design and

Awareness
CT Usage
Research
Problem

CLOs
Analysis

PSO 1

PSO 2

PSO 3
Skills

Outcome 1 3 3 3 - - - - - - - - - 3 2
Outcome 2 3 3 3 2 - - - - - - - 3 3
Outcome 3 3 3 2 - - - - - - - - - 3 2
Outcome 4 3 3 2 - - - - - - - - - 3 3
Outcome 5 3 3 2 - 2 - - - - - - - 3 3
Outcome 6 2 2 3 3 3 - - - - - - - 2 3
Outcome 7 3 3 1 - - - - - - - - - 3 3
Average 3 3 2 3 2 3 3
Course Unitization Plan Theory
Required
Unit CLOs References
Unit Name Contact
No. Addressed Used
Hours
Unit
Introduction 9
1
Introduction: What is digital image and DIP? History,
1,7 1
Applications of DIP
Key stages of Digital Image processing, 1
1,7 1, 4
Advances in machine vision application domain
Image sampling and quantization, spatial resolution, intensity
1 1
resolution
Relationship between pixels: neighbourhood, adjacency and 1
1 1
connectivity, Path, region boundary
Connected component labelling, Distance measure: Euclidian,
1 1
chess board, city block.
Image acquisition and Pre-processing, Intensity transformations, 1
2 1
spatial filtering
Image enhancement: Introduction, Point Processing- image
1 2, 6 1
negative, log transform, dynamic range compression.
Power law or gamma Transformation, gamma correction 2, 6 1
Piecewise linear transformation: contrast stretching, threshold, bit-
1 2, 6 1
plane slicing
Histogram processing: image histogram, histogram equalization 2, 6 1
Numerical on histogram equalization, histogram specification,
1 2 1
numerical on histogram specification
Spatial filters for smoothing operations: linear filters (average and
weighted average), order statistics (nonlinear) filters: median, 1 2, 6, 7 1
min, max filters.
Spatial filters for sharpening operations: Convolution vs.
correlation, objective (integration, differentiation, application of 1 2, 6 1
sharpening),
First order and second order derivative operators and their
1 2 1
response, Laplacian operator, unsharp masking,
Unit Filtering in the Frequency Domain,
9
2 Image Restoration
Frequency domain approach: low pass filtering, high pass
1 2 1, 2, 3
filtering, Laplacian, high boost filtering.
Image transform and its importance, Fourier transform, 1D FT, 1D
1 2 1, 2, 3
Discrete Fourier Transform (DFT)
2D DFT and its property, Holomorphic filtering 1 2 1, 2, 3
Image restoration: Fundamentals, 1 3 1, 2, 3
Noise models, example images affected with noise 1 3 1, 2, 3
Estimation of noise parameters models 1 3 1, 2
Restoration in presence of noise (Spatial domain techniques):
1 3 1, 2
mean filters, order statistics filters
Adaptive local noise filter, adaptive median filter 1 3 1, 2
Estimation of degradation function: (i) by observation, (ii) by
1 3 1, 2
experimentation (iii) mathematical modelling
Unit
Image Segmentation 9
3
Image segmentation: Fundamentals, point, line detection, 1 5, 6 1
Basic edge detection techniques, Hough transform 1 5, 6 1
Thresholding: Bi-modal and Multi-model Histogram, 1 5 1
Noise effect on thresholding, Illumination effect on image
1 5 1
thresholding
 Basic global thresholding, Optimal thresholding using Otsu’s
1 5 1, 2
method
Multi-spectral thresholding, Region based segmentation. 2 5 1, 2
Region growing, Region splitting and Merging. 2 5 1, 2
Unit Color Image Processing,
9
4 Image Compression
Colour image processing: Fundamentals, motivation, full and
2 5 1
pseudo colour image processing
Components of colour, primary and secondary colours,
1 5 1
tristimulus, chromaticity diagram,
Colour models: RGB, CMY, CMYK, HSI 1 5 1, 3
Colour conversion, numerical on colour conversion 1 5 1, 3
Image compression: Motivation, Applications, Compression ratio 1 4 1, 2
Data redundancy- Coding, Inter-pixel and Psycho-visual
1 4 1, 2
redundancy,
JPEG Coding, Huffman Coding 1 4 1, 2
LPZ coding, arithmetic coding, lossless and lossy predictive
1 4 1, 2
coding
Unit
Image representation and Object Recognition 9
5
Image presentation and description- Introduction, Motivations 2 5 3
Shape features (Region-based shape representation and
descriptors) Area, Euler’s number, eccentricity, Elongatedness, 2 5 3
rectangularity, direction, compactness. moments, covex hull.
Texture features, Color features 1 5 3
Object and Pattern Recognition: Pattern and pattern classes. 1 5 3
Matching, classifier role minimum distance or nearest neighbor
1 5 1, 4
classifier.
Matching by correlation, Optimum statistical classifier 1 5 1, 4
Neural network classifier 1 5 1, 4
Total Contact Hours 45

Course Unitization Plan - Lab

Required
S. CLOs References
Experiment Name Contact
No. Addressed Used
Hours 30
Lab Experiment 1:
Perform the following operations using library functions
a. Read, Display and write any color image in other formats. b.
1. Find RED, GREEN and BLUE plane of the color image. c. Convert 2 1 1
color image to grayscale image and binary image
d. Resize the image by one half and one quarter. i.e. Image rotates
by 45, 90 and 180 degrees.
Lab Experiment 2:
Create black and white images (A) of size 1024x1024. Which
consists of alternative horizontal lines of black and white? Each
line is of size 128. Create black and white images (B) of size
1024x1024. Which consists of alternative vertical lines of black and
white? Each line is of size128. Perform the following operations on
2. Image A and Image B. a. Image addition of A and B 2 1 1
b. Subtraction of A and B
c. Multiplying Images of A and B
d. Create a grayscale image of size 256 x 1024. Intensity of image
should vary sinusoidally.
e. Create a white image of size 256x256, with black box of size
58x58 at centre.
3. Lab Experiment 3: 3 2,3 1
 Develop programs for following intensity transformation operation
on a grayscale image. Collect any gray scale image from any
source. Process that image using these operations.
a. Image negative
b. Log transformation and inverse log transform: s = c log (1+r), c
is a const, r ≥ 0. s is pixel intensity of output image, r is the pixel
intensity of input image. Study the effect of constant c on the
quality of output image.
c. Power law transformation: Study the effect of different values of
Gamma used in this transformation.
d. Contrast stretching
e. Gray level slicing
Lab Experiment 4:
Develop programs for following spatial filtering operations on a
grayscale image.
a. Averaging: Implement averaging filtering operations for different
window sizes and study their effect on the quality of output image.
Write your observations on output image quality.
b. Weighted averaging: Implement weighted averaging filtering
operations for different window sizes and study their effect on the
4. 3 4,5 4
quality of output image. Write your observations on output image
quality.
c. Median filtering: Implement weighted averaging filtering
operations for different window sizes and study their effect on the
quality of output image. Write your observations on output image
quality.
d. Max filtering
e. Min filtering
Lab Experiment 5:
Take a grayscale image and add salt and pepper noise. Write
programs for following operations and observe their outputs a.
Linear smoothing or Image averaging
5. b. Weighted averaging 4 2,6 1
c. Median filtering. Compare the output quality among Image
averaging and median filtering.
d. Max filtering
e. Min filtering
Lab Experiment 6:
Write programs to perform following sharpening operations on a
grayscale image
a. Laplacian filter
6. b. Filtering using composite mask 4 2,6 1
c. Unsharp masking
d. High boost filtering
e. Filtering using first order derivative operators such as sobel and
prewitt mask.
Lab Experiment 7:
Write a program to improve contrast of an image using histogram
equalization. The prototype of the function is as below:
7. histogram_equalisation(input_Image, no_of_bins); The function 3 2 1
should return the enhanced image. Consider two low contrast input
images. Study the nature of the output image quality in each case
by varying the number of bins.
Lab Experiment 8:
Take a low contrast grayscale image (A) and a high contrast gray
8. scale image (B). Write a program to improve the contrast of A with 3 2 1
the help of image B using histogram specification or matching. The
prototype of the function is as below: Histogram_sp(input_Image,
 specified_Iage, no_of_bins); The function should return the
enhanced image.
Lab Experiment 9:
Develop programs to implement frequency domain smoothing
filters (Ideal, Butterworth and Gaussian) and apply these filters on a
grayscale image.
a. Compare/comment on the output of Ideal, Butterworth and
Gaussian Low pass Filters having the same radii (cutoff frequency)
9. value. 3 2 1,2,3
b. Consider a suitable gray scale image and demonstrate the ringing
effect on the output of Ideal low pass frequency domain filter.
c. Compare the output of Butterworth low pass filters (order n=2)
for different cutoff frequencies (5, 15, 30, 90, 120).
d. Compare the output of Gaussian low pass filters for different cut-
off frequencies (5, 15, 30, 90, and 120).
Lab Experiment 10:
Develop programs to implement frequency domain
sharpening/High pass filters (Ideal, Butterworth and Gaussian) and
apply these filters on a grayscale image.
a. Compare/comment on the output of Ideal, Butterworth and
Gaussian High pass Filters having the same radii (cutoff frequency)
10. value. 3 2 1,2,3
b. Consider a suitable gray scale image and demonstrate the ringing
effect on the output of Ideal high pass frequency domain filter.
c. Compare the output of Butterworth high pass filters (order n=2)
for different cut-off frequencies (5, 15, 30, 90, 120).
d. Compare the output of Gaussian high pass filters for different
cut-off frequencies (5, 15, 30, 90, and 120).
Total Contact Hours 30

Learning Assessment
Continuous Learning Assessments (50%) End Semester Exam
Bloom’s Level of (50%)
Practical
Cognitive Task CLA-1 CLA2 CLA-3 Mid Th Prac
(20%)
(5%) (5%) (5%) (15%) (30%) (20%)
Remember
Level 1 70% 50% 40% 40% 20% 40% 30%
Understand
Apply
Level 2 30% 50% 60% 40% 30% 40% 30%
Analyse
Evaluate
Level 3 20% 50% 20% 40%
Create
Total 100% 100% 100% 100% 100% 100% 100%

Recommended Resources
1. Gonzalez, R. C. (2009). Digital image processing. Pearson education India.
2. Sridhar, S. (2016) Digital Image Processing, Oxford University Press.
3. Sonka, M., Hlavac, V., & Boyle, R. (2013). Image processing, analysis and machine vision. Springer.
4. Forsyth, D. A., & Ponce, J. (2002). Computer vision: a modern approach. prentice hall professional technical reference.

Other Resources
1. No Data

Course Designers
1. No Data