Wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

CV_6

Total questions: 100

Worksheet time: 50mins

Name
Class
Date
1.

Increasing sensor size while keeping resolution constant mainly improves

a)

Spatial resolution

b)

Signal-to-noise ratio

c)

Quantization levels

d)

Sampling rate

2.

A camera produces blur even for static scenes mainly due to

a)

Low quantization

b)

Long exposure time

c)

Aliasing

d)

Histogram equalization

3.

Brightness constancy assumption is violated when there is

a)

Object rotation

b)

Uniform illumination

c)

Sudden lighting change

d)

Camera translation

4.

If sampling frequency is doubled, Nyquist frequency

a)

Halves

b)

Doubles

c)

Becomes zero

d)

Remains same

5.

Which affects image depth but not spatial resolution?

a)

Sampling

b)

Quantization

c)

Filtering

d)

Scaling

6.

Improving local contrast without affecting global brightness is achieved by

a)

Histogram equalization

b)

Adaptive histogram equalization

c)

Mean filtering

d)

Log transformation

7.

A filter that removes noise but destroys thin lines is likely

a)

Median filter

b)

Gaussian filter

c)

High-pass filter

d)

Laplacian filter

8.

Best noise reduction while preserving edges is achieved using

a)

Mean filter

b)

Median filter

c)

Ideal LPF

d)

Box filter

9.

Negative image transformation is useful for

a)

Edge detection

b)

Highlighting dark details

c)

Noise removal

d)

Compression

10.

Laplacian sharpening amplifies noise because it

a)

Uses first derivative

b)

Works in frequency domain

c)

Emphasizes high frequencies

d)

Reduces contrast

11.

Sharp cutoffs in frequency filters cause

a)

Blur

b)

Ringing artifacts

c)

Quantization

d)

Thresholding

12.

Gaussian filters are preferred since they

a)

Are ideal

b)

Have infinite support

c)

Avoid ringing

d)

Remove edges

13.

Low-pass filtering corresponds to

a)

Differentiation

b)

Integration

c)

Smoothing

d)

Sharpening

14.

Periodic noise removal requires eliminating

a)

DC component

b)

High frequencies

c)

Specific frequency spikes

d)

Spatial outliers

15.

Transform giving both spatial and frequency localization

a)

Fourier

b)

DCT

c)

Wavelet

d)

PCA

16.

Edge detection relies on detecting

a)

Similarity

b)

Discontinuity

c)

Texture

d)

Color

17.

Sobel performs better than Roberts in noise because it

a)

Uses smaller mask

b)

Includes smoothing

c)

Uses second derivative

d)

Works in frequency domain

18.

Histogram-based thresholding assumes

a)

Uniform noise

b)

Bimodal distribution

c)

High resolution

d)

Uniform illumination

19.

Region growing fails mainly due to

a)

Sharp edges

b)

Noise causing false similarity

c)

Strong gradients

d)

Over-filtering

20.

Boundary detection is essential for

a)

Compression

b)

Shape analysis

c)

Noise removal

d)

Sampling

21.

Corners are generally invariant to

a)

Noise

b)

Rotation

c)

Scale

d)

Illumination

22.

Harris corner detection uses

a)

Histogram

b)

Eigenvalues of structure tensor

c)

Hough voting

d)

PCA

23.

Hough Transform complexity increases with

a)

Image size

b)

Parameter space dimensionality

c)

Noise reduction

d)

Thresholding

24.

Feature matching degrades mainly due to

a)

Translation

b)

Viewpoint change

c)

Identical features

d)

High contrast

25.

PCA is unsuitable for

a)

Dimensionality reduction

b)

Feature decorrelation

c)

Non-linear manifolds

26.

Speed of Viola–Jones is due to

a)

CNNs

b)

Integral images

c)

PCA

d)

Optical flow

27.

YOLO formulates detection as

a)

Classification

b)

Segmentation

c)

Regression

d)

Clustering

28.

False positives increase due to

a)

High resolution

b)

Background clutter

c)

Preprocessing

d)

Normalization

29.

Deep learning generalizes poorly when

a)

Data is diverse

b)

Training data is limited

c)

Features are learned

d)

GPU is used

30.

GMM models background using

a)

Uniform distribution

b)

Gaussian mixtures

c)

Binary values

d)

Deterministic values

31.

Optical flow fails when

a)

Texture exists

b)

Motion is slow

c)

Brightness changes

d)

Camera is static

32.

Aperture problem occurs due to

a)

Large motion

b)

Small observation window

c)

Noise

d)

Low resolution

33.

Stereo correspondence matches

a)

Pixels across frames

b)

Pixels across cameras

c)

Objects across scenes

d)

Features across scales

34.

Increasing stereo baseline improves

a)

Matching speed

b)

Depth accuracy

c)

Image quality

d)

Noise suppression

35.

Motion estimation improves with

a)

Large window always

b)

Multi-scale analysis

c)

High noise

d)

Uniform regions

36.

Gaussian blur before Canny is used to

a)

Increase edges

b)

Reduce noise

c)

Improve contrast

d)

Increase resolution

37.

Face recognition fails mainly due to

a)

High resolution

b)

Pose and illumination changes

c)

Noise removal

d)

Edge detection

38.

OCR preprocessing includes

a)

Thresholding & normalization

b)

Feature matching

c)

Stereo vision

d)

Optical flow

39.

Real-time video analytics prioritizes

a)

Accuracy only

b)

Speed–accuracy trade-off

c)

Memory only

d)

Storage

40.

Background subtraction fails when

a)

Camera is fixed

b)

Lighting varies rapidly

41.

Histogram equalization is avoided in medical images because it

a)

Increases contrast

b)

Distorts diagnostic information

c)

Is slow

d)

Reduces resolution

42.

CNNs remove the need for

a)

Training

b)

Manual feature design

c)

Large datasets

d)

GPUs

43.

Multi-scale processing is required because

a)

Objects vary in size

b)

Images are noisy

c)

Quantization is low

d)

Sampling is fixed

44.

Feature normalization improves

a)

Noise

b)

Matching robustness

c)

Edge strength

d)

Segmentation

45.

Tracking fails mainly due to

a)

Object motion

b)

Occlusion

c)

High frame rate

d)

Static background

46.

Grayscale images are used mainly to

a)

Remove information

b)

Reduce computation

c)

Increase noise

d)

Improve color

47.

Edge-based segmentation fails when

a)

Edges are strong

b)

Edges are weak or noisy

c)

Objects are separated

d)

Contrast is high

48.

Deep learning needs large datasets because

a)

Models are shallow

b)

High number of parameters

c)

Low computation

d)

Simple features

49.

Optical flow is expensive because it involves

a)

Pixel-wise motion estimation

b)

Enhancement

c)

Histogram calculation

d)

Thresholding

50.

A complete computer vision pipeline ends with

a)

Enhancement

b)

Segmentation

c)

Recognition

d)

Sampling

51.

A satellite image shows blocky artifacts when zoomed in. The most likely cause is

a)

Noise

b)

Low spatial sampling

c)

High quantization

d)

Histogram equalization

52.

A night-vision camera produces grainy images despite high resolution. The issue is mainly due to

a)

Quantization

b)

Low signal-to-noise ratio

c)

Aliasing

d)

Over-sampling

53.

Two cameras capture the same scene; one image looks darker though resolution is same. The difference is due to

a)

Sampling

b)

Illumination and sensor response

c)

Quantization

d)

Filtering

54.

Medical images require subtle intensity differences to be preserved. Which should be maximized?

a)

Sampling

b)

Quantization levels

c)

Filtering

d)

Compression

55.

A drone image shows distorted object sizes at edges. This is caused by

a)

Noise

b)

Perspective projection

c)

Thresholding

d)

Quantization

56.

A foggy road image needs visibility improvement without amplifying noise. Best method is

a)

Histogram equalization

b)

Contrast stretching

c)

Laplacian sharpening

d)

Thresholding

57.

CCTV footage contains salt-and-pepper noise. The best filter is

a)

Mean

b)

Gaussian

c)

Median

d)

High-pass

58.

After smoothing, fine cracks in a material image disappear. The cause is

a)

Under-sampling

b)

Over-smoothing

c)

Quantization

d)

Aliasing

59.

An image looks washed out after enhancement. The incorrect step is

a)

Normalization

b)

Over-equalization

c)

Smoothing

d)

Conversion to grayscale

60.

Dark objects in X-ray images need emphasis. Suitable transform is

a)

Linear

b)

Logarithmic

c)

Negative

d)

Threshold

61.

Sharp ringing appears after filtering an image. The filter used was likely

a)

Gaussian

b)

Butterworth

c)

Ideal filter

d)

Median

62.

Removing repeating stripes in scanned documents requires

a)

Low-pass filter

b)

High-pass filter

c)

Notch filter

d)

Mean filter

63.

An engineer chooses frequency-domain filtering because

a)

It is local

b)

It handles global patterns efficiently

c)

It removes all noise

d)

It avoids transforms

64.

Excessive high-pass filtering results in

a)

Smoothing

b)

Noise amplification

c)

Contrast loss

d)

Segmentation

65.

A filter with smooth transition avoids ringing because it

a)

Removes DC component

b)

Reduces abrupt frequency changes

c)

Works spatially

d)

Uses median

66.

A road image under shadows fails thresholding. The reason is

a)

Noise

b)

Non-uniform illumination

c)

High contrast

d)

Over-sampling

67.

Broken lane markings are detected as separate edges. Which step is missing?

a)

Smoothing

b)

Thresholding

c)

Edge linking

d)

PCA

68.

Region growing merges background with object due to

a)

Strong edges

b)

Incorrect similarity criteria

c)

High resolution

d)

Sharpening

69.

Watershed segmentation produces too many regions because of

a)

Blur

b)

Over-segmentation

c)

Low resolution

d)

Histogram equalization

70.

For detecting object boundaries in noisy images, best approach is

a)

Thresholding

b)

Canny edge detection

c)

Region growing

d)

PCA

71.

A corner detector fails when object size changes. The detector lacks

a)

Rotation invariance

b)

Scale invariance

c)

Translation invariance

d)

Noise resistance

72.

Harris detector identifies points where

a)

Intensity is maximum

b)

Gradient changes in all directions

c)

Edges are strongest

d)

Histogram peaks

73.

A line is detected even when broken. The algorithm used is

a)

Sobel

b)

Canny

c)

Hough Transform

d)

PCA

74.

Feature matching fails under illumination change because

a)

Features are local

b)

Intensity values vary

c)

Edges disappear

d)

Sampling changes

75.

PCA fails when applied to highly nonlinear image data because it

a)

Is slow

b)

Is a linear technique

c)

Removes noise

d)

Needs labels

76.

Face detection works fast but only for frontal faces. Algorithm used is

a)

CNN

b)

Viola-Jones

c)

YOLO

d)

SIFT

77.

A real-time system detects multiple objects in one pass. The model used is

a)

R-CNN

b)

Fast R-CNN

c)

YOLO

d)

Viola-Jones

78.

Background subtraction fails when lights switch ON/OFF because

a)

Motion stops

b)

Background model becomes invalid

c)

Objects disappear

d)

Frame rate increases

79.

Deep learning detector performs poorly on new environments due to

a)

Over-sampling

b)

Overfitting

c)

PCA

d)

Edge detection

80.

GMM adapts to slow background changes because it

a)

Is static

b)

Updates probability distributions

c)

Uses edges

d)

Uses thresholding

81.

Optical flow fails on a blank wall because of

a)

Noise

b)

Lack of texture

c)

High resolution

d)

Frame rate

82.

An object moves perpendicular to its edge but motion is undetected due to

a)

Noise

b)

Aperture problem

c)

Blur

d)

Quantization

83.

A face detector works only for upright frontal faces but runs very fast. Algorithm used is

a)

CNN

b)

YOLO

c)

Viola-Jones

d)

SIFT

84.

A detector identifies all objects in one forward pass. This indicates use of

a)

Sliding window

b)

R-CNN

c)

YOLO

d)

Hough Transform

85.

Background subtraction fails when sunlight intensity changes suddenly because

a)

Objects stop moving

b)

Background model becomes outdated

c)

Noise increases

d)

Frame rate drops

86.

A deep learning model performs well on training data but poorly on test data due to

a)

Underfitting

b)

Overfitting

c)

Quantization

d)

Sampling

87.

GMM adapts background by

a)

Fixed thresholds

b)

Updating Gaussian parameters

c)

Edge detection

d)

PCA

88.

Optical flow fails on a plain white wall because

a)

Noise

b)

Lack of texture

c)

High resolution

d)

High frame rate

89.

Motion is detected incorrectly along edges due to

a)

Blur

b)

Aperture problem

c)

Noise

d)

Quantization

90.

Stereo vision estimates depth using

a)

Motion

b)

Disparity between views

c)

Histogram

d)

PCA

91.

Increasing baseline improves depth accuracy but also

a)

Reduces noise

b)

Increases correspondence difficulty

c)

Improves illumination

d)

Reduces computation

92.

Motion blur mainly reduces accuracy of

a)

Segmentation

b)

Optical flow

c)

Sampling

d)

Quantization

93.

Converting images to grayscale before processing mainly helps to

a)

Improve color

b)

Reduce computational load

c)

Increase noise

d)

Enhance texture

94.

OCR struggles with cursive handwriting because of

a)

High resolution

b)

Large shape variability

c)

Good contrast

d)

Edge clarity

95.

Facial expression recognition requires analysis of

a)

Color only

b)

Texture and facial muscle movement

c)

Histogram only

d)

Thresholding

96.

Real-time vision systems balance speed and accuracy due to

a)

Storage limits

b)

Computational constraints

c)

Noise

d)

Sampling

97.

Object tracking fails when two objects cross paths due to

a)

Blur

b)

Identity ambiguity

c)

Noise

d)

Resolution

98.

Preprocessing is essential before segmentation because it

a)

Adds features

b)

Improves segmentation reliability

c)

Reduces image size

d)

Performs classification

99.

CNNs outperform handcrafted features because they

a)

Use PCA

b)

Learn hierarchical representations

c)

Avoid training

d)

Use thresholding

100.

Multi-scale processing is required because

a)

Images are noisy

b)

Objects appear at different sizes

c)

Sampling varies

d)

Quantization changes

Similar Resources on Wayground

Content Section Worksheet Questions (Grade 1 adapted MCQs)

100 questions

Content Section Worksheet Questions (Grade 1 adapted MCQs)

quiz

Worksheet

•

1st Grade