Image Processing

Course Title: Image Processing

Course No: CSC321

Nature of the Course: Theory + Lab

Semester: V

Full Marks: 60 + 20 + 20

Pass Marks: 24 + 8 + 8

Credit Hrs: 3

Course Description

This course covers the investigation, creation and manipulation of digital images by computer. The course consists of theoretical material introducing the mathematics of images and imaging. Topics include representation of two-dimensional data, time and frequency domain representations, filtering and enhancement, the Fourier transform, convolution, interpolation. The student will become familiar with Image Enhancement, Image Restoration, Image Compression, Morphological Image Processing, Image Segmentation, Representation and Description, and Object Recognition.

Course Objectives

The objective of this course is to make students able to:

  • Develop a theoretical foundation of Digital Image Processing concepts.
  • Provide mathematical foundations for digital manipulation of images; image acquisition; preprocessing; segmentation; Fourier domain processing; and compression.
  • Gain experience and practical techniques to write programs for digital manipulation of images; image acquisition; pre-processing; segmentation; Fourier domain processing; and compression.
Course Contents
Unit 1: Introduction (5 Hrs.)
  • Digital Image: Definition of digital image, pixels, representation of digital image in spatial domain as well as in matrix form. (1 hr)
  • Fundamental steps in Image Processing: Block diagram of fundamental steps in digital image processing, application of digital image processing system, Elements of Digital Image Processing systems. (1 hr)
  • Element of perception: Structure of the Human, Image Formation in the Eye, Brightness Adaptation and Discrimination. (1 hr)
  • Sampling and Quantization: Basic Concepts in Sampling and Quantization, Representing Digital Images, Spatial and Gray-Level Resolution. (1 hr)
  • Some basic relationships like Neighbors: Neighbors of a Pixel, Adjacency, Connectivity, Regions, and Boundaries, Distance Measures between pixels. (1 hr)
Unit 2: Image Enhancement and Filter in Spatial Domain (8 Hrs.)
  • Basic Gray Level Transformations: Point operations, Contrast stretching, clipping and thresholding, digital negative, intensity level slicing, log transformation, power law transformation, bit plane slicing. (2 hrs)
  • Histogram Processing: Unnormalized and Normalized Histogram, Histogram Equalization, Use of Histogram Statistics for Image Enhancement. (1 hr)
  • Spatial operations: Basics of Spatial Filtering, Linear filters, Spatial Low pass smoothing filters, Averaging, Weighted Averaging, Non-Linear filters, Median filter, Maximum and Minimum filters, Midpoint and Alpha trimmed mean filters; High pass sharpening filters, High boost filter, high frequency emphasis filter, Gradient based filters, Robert Cross Gradient Operators, Prewitt filters, Sobel filters, Second Derivative filters, Laplacian filters. (4 hrs)
  • Magnification by replication and interpolation. (1 hr)
Unit 3: Properties of Fourier Transform (8 Hrs.)
  • Introduction: Introduction to Fourier Transform and the frequency Domain, 1-D and 2-D Continuous Fourier transform, 1-D and 2-D Discrete Fourier transform. (1 hr)
  • Fast Fourier Transform: Computing and Visualizing the 2D DFT (Time Complexity of DFT), Derivation of 1-D Fast Fourier Transform, Time Complexity of FFT, Concept of Convolution, Correlation and Padding. (2 hrs)
  • Other Transforms: Hadamard transform, Haar transform and Discrete Cosine transform. (2 hrs)
  • Smoothing Frequency Domain Filters: Ideal Low Pass Filter, Butterworth Low Pass Filter, Gaussian Low Pass Filter. (1 hr)
  • Sharpening Frequency Domain Filters: Ideal High Pass Filter, Butterworth High Pass Filter, Gaussian High Pass Filter, Laplacian Filter. (2 hrs)
Unit 4: Image Restoration and Compression (8 Hrs.)
  • Image Restoration: Introduction, Models for Image degradation and restoration process, Noise Models (Gaussian, Rayleigh, Erlang, Exponential, Uniform and Impulse), Estimation of Noise Parameters. (2 hrs)
  • Restoration Filters: Mean Filters: Arithmetic, Geometric, Harmonic and Contraharmonic Mean Filters; Order Statistics Filters: Median, Min and Max, Midpoint and Alpha trimmed mean filters; Band pass and Band Reject filters: Ideal, Butterworth and Gaussian Band pass and Band Reject filters. (2 hrs)
  • Image Compression: Introduction, Definition of Compression Ratio, Relative Data Redundancy, Average Length of Code; Redundancies in Image: Coding Redundancy (Huffman Coding), Interpixel Redundancy (Run Length Coding) and Psychovisual Redundancy (4-bit Improved Gray Scale Coding: IGS Coding Scheme). (4 hrs)
Unit 5: Morphological Operations (2 Hrs.)
  • Introduction to Morphological Image Processing: Definition of Fit and Hit. (1 hr)
  • Dilation and Erosion, Opening and Closing. (1 hr)
Unit 6: Image Segmentation (8 Hrs.)
  • Discontinuity Based Techniques: Point Detection, Line Detection, Edge Detection using Gradient and Laplacian Filters, Mexican Hat Filters, Edge Linking and Boundary Detection, Hough Transform. (3 hrs)
  • Similarity Based Techniques: Thresholding: Global, Local and Adaptive; Region Based Segmentation: Region Growing Algorithm, Region Split and Merge Algorithm. (4 hrs)
Unit 7: Representation and Description (5 Hrs.)
  • Representations, Description and Recognition: Introduction to some descriptors: Chain codes, Signatures, Shape Numbers, Fourier Descriptors. (2 hrs)
  • Pattern Recognition: Overview of Pattern Recognition with block diagram; Patterns and pattern classes, Decision-Theoretic Methods, Introduction to Neural Networks and Neural Network based Image Recognition. (3 hrs)
Laboratory Works

Students are required to develop programs in related topics using suitable programming languages such as MatLab or Python or other similar programming languages.

Text Books
  • Rafael C. Gonzalez and Richard E. Woods, “Digital Image Processing”, Pearson Edition, Latest Edition.
Reference Books
  • I. Pitas, "Digital Image Processing Algorithms", Prentice Hall, Latest Edition.
  • A. K. Jain, “Fundamentals of Digital Image Processing”, Prentice Hall of India Pvt. Ltd., Latest Edition.
  • K. Castlemann, “Digital Image Processing”, Prentice Hall of India Pvt. Ltd., Latest Edition.
  • P. Monique and M. Dekker, “Fundamentals of Pattern Recognition”, Latest Edition.

Model Question

Section A

Attempt any two questions. (2 × 10 = 20)

  1. Differentiation between Discrete Fourier Transform (DFT) and Fast Fourier Transform (FFT). Explain the FFT algorithm for one-dimensional case. [3+7=10]

  2. Given the following frequency table obtained from the histogram of a 16 X 16, 8 level image. 
                                M   0  1   2  3  4  5  6  7
                            Nm  15 6  70 16 31 35 32 51
    Determine:
    • The mean and variance of the image intensity levels. [5]
    • Image Histogram Equalization. [5]

  3. Write a brief note on the application of image processing in medical imaging. [10]

  4. Write the algorithm for the median filter. How is it different from the mean filter? [5+5=10]
Section B

Attempt any six questions. (6 × 5 = 30)

  1. What is Histogram Equalization? [5]

  2. Explain the concept of “Image Degradation Model”. [5]

  3. Write a brief note on "Region Growing" technique in image segmentation. [5]

  4. Discuss any two noise models used in image restoration. [5]

  5. Differentiate between "Smoothing" and "Sharpening" filters. [5]

  6. Explain the concept of "Morphological Operations". [5]

  7. Write the procedure for "Adaptive Thresholding". [5]

  8. Explain "Dilation" and "Erosion" operations in morphological image processing. [5]