AI advancements in image processing

2. What is the role of the convolutional layer in a CNN?

• Reduces the number of parameters

• Detects spatial patters in an image

• Converts image to text

• Reduces the number of colors in an image

What is the primary purpose of using Convolutional Neural Networks (CNNs) in image processing

• To reduce the size of an image

• To extract spatial features from an image

• To improve text analysis

• To convert images into grayscale

What does a kernel (or filter) do in a CNN?**

• A) Randomly removes pixels from an image

• B) Scans through the image to extract features

• C) Converts an image into a binary format

• D) Adds extra layers to the CNN

What is the function of the pooling layer in CNNs?

• To increase resolution of an image

• To reduce computational complexity by downsampling

• To add more color channels

• To increase the number of neurons

What is the most common activation function used in CNNs?

• A) Sigmoid

• B) Tanh

• C) ReLU (Rectified Linear Unit)

• D) Softmax

Which technique prevents overfitting in CNNs?**

• A) Increasing the number of filters

• B) Adding more fully connected layers

• C) Using dropout regularization

• D) Increasing the learning rate

What is the role of stride in the convolution operation?**

• A) Controls the step size of the filter movement

• B) Determines the number of channels in an image

• C) Increases the image resolution

• D) Controls the color intensity

Which of the following is an advantage of CNNs over traditional machine learning for image classification?**

• A) They require handcrafted features

• B) They automatically learn spatial hierarchies of features

• C) They only work with grayscale images

• D) They do not require large datasets

What is the main advantage of using batch normalization in CNNs?**

• A) Reduces the number of parameters

• B) Normalizes the inputs to stabilize and accelerate training

• C) Converts images to black and white

• D) Increases model complexity

What is the main purpose of transfer learning in CNNs?

• A)To reduce the depth of the network

• B) To use pre-trained models for new tasks with limited data

• C) To generate synthetic images

• D) To remove noise from images