Block 2: Image Recognition With CNNs Flashcards
What is the advantage of using CNNs over traditional image processing techniques?
CNNs are more effective in capturing spatial hierarchies and features in images, leading to better performance in image-related tasks.
Name one key component of a CNN.
One key component of a CNN is the convolutional layer.
Explain the role of the pooling layer in a CNN.
The pooling layer reduces the spatial size of the representation, decreasing the number of parameters and computation in the network.
How do convolutional layers work in a CNN?
Convolutional layers apply filters to the input to create feature maps that capture spatial features.
What are the typical uses of fully connected layers in a CNN?
Fully connected layers combine features learned by the convolutional layers to make predictions or classifications.
Describe the importance of activation functions in CNNs.
Activation functions introduce non-linearity in the network, allowing it to learn complex patterns.
How does a CNN differ from a standard neural network in terms of architecture?
A CNN has specialized layers like convolutional and pooling layers designed for processing grid-like data, unlike standard neural networks.
Discuss the challenges in training deep CNNs.
Challenges include the need for large datasets, high computational power, and the risk of overfitting.
Explain how CNNs can be used in tasks beyond image recognition.
CNNs can be adapted for tasks like video analysis, natural language processing, and complex pattern recognition in various data types.
Analyze the impact of different kernel sizes in convolutional layers.
Different kernel sizes can capture various levels of details in images; larger kernels capture broader features while smaller kernels capture finer details.
Describe the process of backpropagation in CNNs.
Backpropagation in CNNs involves updating filter weights to minimize the loss, similar to standard neural networks, but considering the spatial hierarchy.
How do CNNs contribute to advancements in computer vision?
CNNs have greatly advanced computer vision by improving accuracy and efficiency in tasks like object detection, facial recognition, and image classification.
What is the purpose of a filter in a CNN?
A filter in a CNN is used to extract specific features from the input image, like edges, textures, or colors.
Explain the concept of stride in a CNN.
Stride refers to the number of pixels by which the filter moves across the input image. A larger stride reduces the spatial dimension of the output feature map.
How do padding techniques like ‘same’ and ‘valid’ padding differ in CNNs?
‘Same’ padding adds zeros around the input image to keep the output size the same as the input, while ‘valid’ padding does not add padding and reduces the output size.
Discuss the role of data augmentation in training CNNs.
Data augmentation involves artificially increasing the diversity of the training dataset by applying random transformations to images. It helps improve the robustness and generalization of CNNs.
Explain the concept of feature map in CNNs and its importance.
A feature map is the output generated from each filter in a CNN. It represents the features detected by the filter, such as edges or textures, and is crucial for learning complex patterns in the data.
How do CNN architectures like AlexNet and VGG differ in terms of design and performance?
AlexNet and VGG have different layer configurations, with VGG having a deeper structure with more convolutional layers. VGG is known for its simplicity and depth, while AlexNet was pioneering in using ReLU and dropout.
