Scribd
Upload
455 views2 pages

Quiz Solution

This 3 sentence summary provides the high level information about the document: The document appears to be a quiz for a computer science or engineering course, containing 15 multiple choice questions about topics related to digital image processing and computer vision, such as color models, edge detection techniques, histogram equalization, and the Hough transform. The questions cover concepts like gradient operators, thresholding, interpolation, and segmentation. Students are instructed to provide the letter of the correct multiple choice answer for each question.

Uploaded by

VIKAS AGNIHOTRI
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
455 views2 pages

Quiz Solution

This 3 sentence summary provides the high level information about the document: The document appears to be a quiz for a computer science or engineering course, containing 15 multiple choice questions about topics related to digital image processing and computer vision, such as color models, edge detection techniques, histogram equalization, and the Hough transform. The questions cover concepts like gradient operators, thresholding, interpolation, and segmentation. Students are instructed to provide the letter of the correct multiple choice answer for each question.

Uploaded by

VIKAS AGNIHOTRI
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Quiz 4 CSE340/540 ECE 350

Total Marks # 8, Duration # 40 minutes


Instructions :
- For objective/ true-false, just mention the correct option or True/False
- Symbols have their usual meanings.
Q1. For histogram processing, HSI is a better color model because we can perform equalization on [.5]

I. I component II. Both S and I components III. Both H and I components IV All components

Q2. If the Hue component is changed by 360 degrees, the resultant RGB image would look more of [.5]

I. Red II. Magenta III. No change IV. None of the above

Q3. For a given 3x3 magnitude gradient G(x,y), suppose hysterisis thresholding gives the following matrices (left one
is obtained after high threshold and right after low threshold)

0 0 255 0 0 255

100 0 200 0 0 200


0 0 235 50 0 235

Will the value 50 be considered as belonging to edge pixel? T/F. [.5]

Q4. Which of the following cannot be used as a gradient operator [.5]

I. Prewitt II. Laplacian III. Sobel IV. Roberts

Q5. If an image is interpolated and then edge detection is performed, the best results will be in case of [1]

I. nearest neighbor, LoG II. Bilinear, Sobel III. Bicubic, Canny IV. Bicubic, Prewitt

Q6. Suppose a 2x2 matrix is to be segmented using Otsu’ scheme. Let the four pixel values be {0, 0, 0, 1}. Probability
of class1, and Class conditional mean for class 1 and class 2 for a threshold of 1 (all pixels equal to and below the
threshold value will form one class) will be respectively given by [2]

I. (1, 1/4, 0) II. (1, 3/4, 0) III. (1, 0, 1) IV. (3/4, 1/4, 1)

Q15. Two images A and B have prominent lines oriented at finite slopes 𝑚𝐴 and 𝑚𝐵 , where 𝑚𝐵 < 0.2. When a Hough
transform is performed, one can observe a vote at accumulator cell for Image A 𝒜(𝑚𝐴 , 𝑐𝐴 ) and
a vote at accumulator cell for image B 𝒜(1 − 𝑚𝐵 , 𝑐𝐵 ). 𝑐𝐴 and 𝑐𝐵 are the corresponding intercepts for slopes
𝑚𝐴 and 𝑚𝐵 . 𝒜(𝑚𝐴 , 𝑐𝐴 ) simply denotes a cell in slope-intercept axis with the center at (𝑚𝐴 , 𝑐𝐴 ).

i. low, low ii. Low, high iii. High, low iv high, high [1]

Q8. A non-max suppression is to be done on the following 3x3 pixel neighborhood –


X
0 0 255

0 255 255
255 255 255

Y
Find out and draw the direction along which you will compute non-max suppression. Highlight the neighbors against
which non-max suppression is performed for the center pixel. [2]
Ans. Applying x and y gradient operators at the center element, 𝑔𝑥 = −0 + 255, 𝑔𝑦 = −0 + 255 [1]

𝑔𝑦
And gradient direction 𝛼 = 𝑎𝑡𝑎𝑛 (𝑔𝑥 ) = 𝜋/4 perpendicular to edge. [.5]

edge
0 + 0 255

0 255 255
255 255 255
+
Gradient direction

Marked pixels with '+' are used for non-max suppression. The response at center pixel is compared with these two
pixels. [.5]

You might also like