5. IMAGE AND VIDEO ANALYTICS ON EDGE

Question 1

What will be the total size of a HoG feature vector if a 9 bin approach based on 8x8 window which is normalized for 16x16 block?

a. 256 x 1
b. 36 x 1
c. 9 x 1
d. 8 x 1

Answer 1

36 x 1

Question 2

Given an extremely resource constrained edge device (e.g. Raspberry Pi Pico) which is the following can be employed for hand land mark detection?

a. A convolutional neural network (CNN) model
b. A fully connected neural network model
c. Image features + support vector machine (SVM)
d. Any of the above

Answer 2

Image features + support vector machine (SVM)

Question 3

In a gray scale image which is uint8 type, what will be the value corresponding to gray color?

a. 0
b. 255
c. 124
d. None of the above

Answer 3

124

Question 4

What does image gradient infer?

a. Low frequency in the image
b. Smooth texture in the image
c. Colors in the image
d. All sudden transitions (edges) in the image

Answer 4

All sudden transitions (edges) in the image

Question 5

From the matrix size below, identify a color image

a. 256 x 256 x 3
b. 256 x 256
c. 256 x 256 x 2
d. 256 x 245 x 5

Answer 5

256 x 256 x 3

Question 6

Which of the following is useful for extracting depth from a video?

a. Local binary descriptor
b. Edges
c. Optical flow estimations
d. All of the above

Answer 6

Optical flow estimations

Question 6

Which of the following is the definition of histogram of an image?

a. A measure of frequency of occurrence of image intensities
b. A measure of frequency of occurrence of edges in the image
c. A measure of noise in the image
d. All of the above

Answer 6

A measure of frequency of occurrence of image intensities

Question 7

Select a suitable window size for obtaining large image gradients in the image of size 256 x 256

a. 3 x 3
b. 8 x 8
c. 64 x 64
d. 256 x 256

Answer 7

64 x 64

Question 8

Which is supposed to be a high frequency component in an image?

a. Random noise
b. Edges in image
c. Both a and b
d. None of the above

Answer 8

Both a and b

Question 9

What is the preferred algorithm for hand gesture based interface?

a. Direct hand landmark detection – gesture control
b. Hand detection – Hand landmark detection – gesture control
c. Direct gesture control
d. None of the above

Answer 9

Hand detection – Hand landmark detection – gesture control

Question 10

What is the starting point of any edge detection?

a. Smoothing followed by gradient operation
b. Gradient operation alone
c. Gradient followed by smoothing operation
d. All of the above

Answer 10

Smoothing followed by gradient operation

Question 11

Which of the following feature(s) can be used for face detection in the image on an edge device?

a. Histogram of oriented gradients
b. Histogram of gradients
c. Histogram of oriented textures
d. None of the above

Answer 11

Histogram of oriented gradients

Question 12

Which of the following images has the highest resolution?

a. 256 x 256
b. 2048 x 2048
c. 1024 x 1024
d. 16 x 16

Answer 12

2048 x 2048

Question 13

Which is preferable edge only image analytics?

a. Image segmentation
b. Fingerprint informatics
c. Object detection
d. All of the above

Answer 13

Fingerprint informatics

Question 14

Which is the following is a reason for increased popularity for computer vision on the edge?

a. Low latency is critical for fast action (e.g. motion segmentation)
b. Sensitivity of the data is critical for many applications
c. Redundancy of the data can be reduced (video summarization)
d. All of the above

Answer 14

All of the above