Quiz

[Dinesh Pandey]

1: Gaussian blurring and other filters act like erasers on the noisy pixels by recalculating their pixel values based on neighboring pixels.
Answer: True

2: The input to a Convolutional Neural Network will always be a 2-D array.
Answer: False

3: Computer Vision" includes which of the following tasks
Answer: Video analysis
        Image classification
        Video analysis
        Similarity learning

4: Out of the following images of same dimensions(pixel size), which one will take the least storage space?
Answer: Black and White

5: Internal layers of Convolutional Neural Networks (CNNs) can be considered as:
Answer: mage filters

6:
Consider the below statements:

I. Artificial neural network takes a vector as input. Spatial structure is ignored.

II. Just like ANN, which work on vectors, CNN also works on vectors. Thus it also ignores spatial structures.

Answer: Only I is correct

7:
An image size is 8X8 pixel. The space required to store this image will be:
Answer: 192 bytes


8: Sets of points in the image which have a strong gradient magnitude are known as:
Answer: Edges

9: The true statements of the following are
Answer: Amplitude quantization is the process of representing the amplitude of the 2D signal at a given coordinate as an integer value.


10: Reasons to apply blurring filters to images are:
Answer:
To avoid overfitting of our model.
To remove unwanted specks of brightness.
To reduce the noise in the image.

[Dinesh Pandey]

Sorry not for this group. Please ignore this email.

Thanks

Dinesh