Wednesday, August 6, 2014

IMAGE PROCESSING ELECTIVES MAY 2007 ELECTRONICS AND TELECOMMUNICATION SEMESTER 8

IMAGE PROCESSING ELECTIVES MAY 2007 ELECTRONICS AND TELECOMMUNICATION SEMESTER 8

Con.3270-07.                              (REVISED COURSE)                                       ND-1522 

                                                          (3 Hours)                                          [Total Marks:-100] 

N.B:- (1) Questions No.1 is compulsory.
          (2) Answer any four out of remaining six questions.
          (3) Figures to the right indicate full marks.
          (4) Assume suitable data if necessary.

1. Justify/contradict following statements:- [20 Marks]
    (a) If kernel of an image transform is separable and symmetric, the transform can be expressed
         in matrix form.
    (b) All image compression techniques are invertible.
    (c) Chain codes can be made invariant to translation and rotation.
    (d) Median filter is best solution to remove salt and pepper noise.
    (e) Walsh transform is nothing but sequency ordered matrix.

2. (a) A 64 x 64 image, represented by 8 bit/pixel has following gray level Distribution. [10 Marks]    
Gray Level
0
1
2
3
4
5
6
7
No. of pixel
128
75
280
416
635
1058
820
684
         Perform histogram equalization and give new distribution of gray level. Show plots of original
         and equalised image.
     (b) Explain basic principles of detecting following in the images: - [10 Marks]
           (i) Points   (ii) Lines   (iii) Edges
          Give 3x3 mask for each of them and explain their operation.

3. (a) Given below is the table of 8 symbols and their frequency of occurrence :-[10 Marks]
Symbol
S₁
S₂
S₃
S₄
S₅
S₆
S₇
S₈
Frequency
0.25
0.15
0.06
0.08
0.21
0.14
0.07
0.04
         Give Huffman code for each eight symbol
    (b) Calculate entropy of the source. [4 Marks]
    (c) Find its average length. [3 Marks]
    (d) Calculate efficiency of the code obtained in 3(a) . [3 Marks]

4. (a) Show that 2D-DCT of an image can be computed by row and column passes with
         1D-DCT algorithm. [10 Marks]        
     (b) For 2x2 transform A and the image U. [10 Marks]
          Calculate transformed image V and the basis image.

5. (A) Explain the method of segmentation of images by region splitting and merging. [10 Marks]
    (b) Explain the following morphological operation in terms of dilation and erosion operations Hit
         or Miss, Open , Close, Boundary ,Skeleton. [10 Marks]

6. (a) Explain different types of color models and where they use. How do convert colors from
         RGB to HSI model? [10 Marks]
    (b) Assuming that edge starts in the first row and ends in the last row, for following gray-level
          image, sketch all possible paths an determine edge corresponding to minimum coast
          path. [10 Marks]

7. Write short notes on(any four):- [20 Marks]
     (a) Fourier descriptors
     (b) DPCM/DCT Encoder
     (c) 2D-wavelet transform filterbank
     (d) Image Compression Standard
     (e) Pseudocolor Image Processing  

No comments:

Post a Comment