CSE509F23MTsoln

pdf

School

Arizona State University *

*We aren’t endorsed by this school

Course

509

Subject

Computer Science

Date

Dec 6, 2023

Type

pdf

Pages

8

Uploaded by SuperHumanOstrich3671

Report
CSE 509 Midterm Exam Wednesday, October 11th, 2023 - 75 minutes - no materials Answer all 20 questions by filling the circle next to your choice. 1. A typical refresh rate for an LCD is: O 50 kHz O 60 Hz O 24 frames/second O 3 GHz 2. Which statement is true? O An LCD makes use of a single polarizing filter. O An LCD makes use of two polarizing filters. O An LCD does not need a backlight. O An LCD makes use of subtractive color mixing. 3. Which statement is true? O Humans are tetrachromats. O A display gamut is typically drawn as a rectangle on the CIE diagram. O YCbCr 4:2:2 refers to a form of chroma subsampling. O RGB encoding is typically used for broadcasting video. 4. One frame of HD video with a resolution of 1920x1080 with 8 bits per color channel (RGB) needs how much memory to store? O 6.2208 MB O 0.9216 MB O 2 MB O 8.2944 MB Print your name inside the box:
5. Sharpening an image will likely cause its histogram to become: O wider O narrower O bimodal O equalized 6. Match the images below with the histograms by selecting the correct order: O a b c d O a c b d O d b c a O b d c a
7. A 3x3 box filter is convolved with an image. Given a portion of the image below, what is the value of the output pixel when the filter is applied to the circled pixel? 0 0 0 0 0 90 90 90 0 90 90 90 0 90 90 90 O 20 O 40 O 60 O 90 8. A weakness of the box filter is that it: O causes aliasing O creates axis-aligned streaks O creates Gaussian noise O increases the sharpness of the image
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help
9. Given the following code: image1 = np.zeros((32,32), np.uint8) image2 = np.zeros((32,32), np.uint8) image1[13:20,13:20] = 255 * np.ones((7,7), np.uint8) / 49.0 image2[11:22,11:22] = 255 * np.ones((11,11), np.uint8) / 121.0 Match the dft magnitude spectra below with image1 and image2: (a) (b) O (a) goes with image1 and (b) goes with image2 O (a) goes with image2 and (b) goes with image1 O The spectra cannot be for either image1 or image2 10. What kind of kernel is this? O a Laplacian O a Gaussian O a Laplacian of a Gaussian O a box filter
11. The Sobel operator is a low-pass filter. O true O false 12. Aliasing can occur when: O The sampling frequency is too high. O The sampling frequency is too low. O The sampling frequency is perfect. O The sampling frequency is zero. 13. Increasing the variance of a Gaussian kernel will increase the variance of its DFT. O true O false
14. Match the images below with the spectra by selecting the correct order: a b c O a b c O b c a O c a b O b a c
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help
15. If there is only a very small shift in the intensity between two images then we expect: O the MSE to be close to zero and the PSNR to be around 1dB O the MSE to be close to zero and the PSNR to be around 40dB O the MSE to be high and the PSNR to be low O None of these. 16. What do you expect the SSIM to be for the case in question 15? O close to 0.0 O close to 1.0 O close to 100.0 O close to 1000.0 17. What type of noise does the following code add to an image? def noise( image, prob ): output = np.zeros(image.shape,np.uint8) thres = 1 - prob for i in range(image.shape[0]): for j in range(image.shape[1]): rdn = random.random() if rdn < prob: output[i][j] = 0 elif rdn > thres: output[i][j] = 255 else: output[i][j] = image[i][j] return output O Gaussian noise O salt-and-pepper noise O structural noise O quantization noise
18. Which component of the SSIM is a measure of contrast similarity? O the top O the middle O the bottom 19. One way to deal with occlusion when doing motion estimation with block matching is to: O use the diamond algorithm O use the three step search algorithm O use both forward and backward motion estimation 20. If we want to do sub-pixel EBMA then we need to do: O bilinear interpolation of the target image O bilateral filtering of the target image O Gaussian filtering of the anchor image O Box filtering of the afterimage