CS-803-A Image Processing and Computer Vision - May 2024 Question Paper

Under each question, How to answer lists the points an answer should cover, outlined by CampusPrep from the syllabus. Check the details against your textbook.

1. 7 Marks
   aExplain the concept of computer vision which is mandatory in image processing.
   How to answer

   • Define computer vision and its goal of scene understanding
   • Relation to image processing: low-level vs high-level vision
   • Key stages: acquisition, processing, segmentation, recognition, interpretation
   • Applications mandating vision in image processing
2. 7 Marks
   bDiscuss various color vision models.
   How to answer

   • Need for color models in vision
   • RGB model: additive primaries
   • CMY/CMYK, HSI/HSV models
   • Comparison and use in processing
3. 7 Marks
   aWhat are the components of digital image processing system? Explain each in detail.
   How to answer

   • Block diagram of digital image processing system
   • Components: acquisition, storage, processing, communication, display
   • Function of each component
   • Example flow of image data
4. 7 Marks
   bDefine briefly the following terms:

   1. i)Image restoration
   2. ii)Compression
   3. iii)Segmentation
   4. iv)Morphological processing
   How to answer

   • One-line definition of each term
   • Purpose in image processing pipeline
   • Distinguishing restoration vs enhancement, compression, segmentation, morphology
   • Example application for each
5. 7 Marks
   aIllustrate how the image is digitized by sampling and quantization process?
   How to answer

   • Continuous image sensing concept
   • Sampling: spatial discretization and resolution effect
   • Quantization: gray-level discretization and bits
   • Resulting digital image matrix representation
6. 7 Marks
   bShow the various techniques in frequency domain to enhance a image with necessary examples.
   How to answer

   • Frequency domain filtering framework with DFT
   • Smoothing: ideal/Butterworth/Gaussian low-pass filters
   • Sharpening: high-pass and homomorphic filtering
   • Examples of enhanced images
7. 7 Marks
   aDistinguish between spatial and frequency domain image enhancement.
   How to answer

   • Definition of spatial vs frequency domain
   • Operation directly on pixels vs on Fourier coefficients
   • Comparison on basis, complexity, filtering
   • Advantages/limitations of each
8. 7 Marks
   bExplain Hough transforms with the help of suitable derivations.
   How to answer

   • Principle of Hough transform for line detection
   • Parametric representation $\rho = x\cos\theta + y\sin\theta$
   • Voting in accumulator array steps
   • Extension to curves and suitable example
9. 7 Marks
   aDiscuss the procedure for conversion from RGB color model to HIS color model.
   How to answer

   • RGB to HSI conversion need
   • Formulas for intensity, saturation, hue
   • Steps: normalize RGB, compute $I$, $S$, $H$
   • Geometric cone interpretation
10. 7 Marks
    bExplain how the line is detected in the image and give the masks that are used to detect it?
    How to answer

    • Principle of second-derivative line detection
    • Four directional $3\times3$ masks for horizontal, vertical, oblique
    • Convolution response and thresholding
    • Example of detected lines
11. 7 Marks
    aExplain the use of neural network structures for pattern recognition with an example.
    How to answer

    • Pattern recognition problem and neural approach
    • Architecture: input, hidden, output layers
    • Training with example patterns
    • Example classification result
12. 7 Marks
    bExplain Boundary descriptors in detail with a neat diagram.
    How to answer

    • Definition of boundary descriptors
    • Fourier descriptors, chain codes, shape numbers
    • Neat diagram of boundary representation
    • Invariance properties
13. 7 Marks
    aExplain the segmentation techniques that are based on finding the regions directly.
    How to answer

    • Region-based vs discontinuity-based segmentation
    • Region growing, splitting and merging
    • Thresholding and watershed approach
    • Advantages for direct region finding
14. 7 Marks
    bExplain inverse perspective projection algorithm.
    How to answer

    • Need for inverse perspective mapping
    • Perspective projection geometry
    • Steps of inverse mapping algorithm
    • Application in 3D vision
15. 14 Marks
    Write brief notes:

    1. a)Feature extraction
    2. b)Knowledge Representation
    3. c)Backtracking algorithm
    How to answer

    • Feature extraction: definition and types of features
    • Knowledge representation: schemes for vision
    • Backtracking algorithm: search principle and use
    • Brief example for each