Convolutional Neural Networks - Filters Deep Dive

1 Convolutional Neural Networks - Introduction to CNNs and Filters

2 CNN MCQ - CNNs allows the network to concentrate on low-level features in the first hidden layer, then assemble them into higher-level features in the next hidden layer, and so on.

3 CNN MCQ - CNNs take 1D vector flattened images as inputs.

4 CNN MCQ - Convolution is the process of combining two signals or the transformation of a signal by a system / filter.

5 CNN MCQ - The distance between two consecutive receptive fields is called the

6 CNN MCQ - Which of the following allows the output to have the same height and width as those of the input?

7 CNN MCQ - Zero Padding is also known as

8 CNN MCQ - No padding is added in a VALID padding

9 CNN MCQ - Convolution neural network need more memory than DNNs

10 CNN MCQ - What is the effect of making CNN's deeper?

11 Convolutional Neural Networks - Filters Deep Dive

12 CNN MCQ - In CNNs, a neuron’s weights can be represented as a small image the size of the receptive field also known as

13 CNN MCQ - Within one feature map, all neurons share the same parameters i.e. weights and bias term, but different feature maps may have different parameters.

14 CNN MCQ - The fact that all neurons in a feature map share the same parameters dramatically reduces the number of parameters in the Model

15 CNN MCQ - A convolutional layer simultaneously applies multiple filters to its inputs, making it capable of detecting multiple features anywhere in its inputs.

16 CNN MCQ - CNN can detect multiple instances of a feature in an image?

17 CNN MCQ - 30 filters of size 3x3 are applied on a feature map of shape 300 * 300 * 10

18 CNN MCQ - A 5x5 filter is applied on an RGB image of size 1028x1028

19 CNN MCQ - Training needs more memory than inference

20 CNN MCQ - CNN don’t use back-propogation during inference

21 Convolutional Neural Networks - Pooling Layers

22 CNN MCQ - Maxpools is used for

23 CNN MCQ - The goal of pooling layers is to

24 CNN MCQ - A pooling layer typically works on every input channel independently. So the output depth is the same as the input depth

25 CNN MCQ - We may also pool over the depth dimension, in which case Image’s spatial dimensions (height and width) remain unchanged, but the number of channels is reduced.

26 Convolutional Neural Networks - Building CNNs

27 Convolutional Neural Networks - CNN Architectures

28 Convolutional Neural Networks - Classification with Keras

29 Convolutional Neural Networks - Transfer Learning with Keras

30 CNN MCQ - Transfer training is faster than training from scratch

31 CNN MCQ - A 1080x1080 black and white image also includes data from a UV channel. The shape of the image will be