Convolutional neural network

A convolutional neural network (CNN or ConvNet) is a specialized type of neural network architecture designed for processing visual imagery.

Key properties of CNNs:

• Use convolutional layers that apply filters to the input
• Translation invariant feature detection
• Hierarchical feature extraction
• Subsampling layers to reduce dimensions
• Fully connected layers at output for classification

CNNs Excel at:

• Image and video recognition
• Object detection and segmentation
• Image classification and labeling
• Visual processing and analysis

Notable CNN architectures include LeNet, AlexNet, VGGNet, Inception, and ResNet. CNNs have become the standard model type for computer vision tasks.

CNNs leverage properties of visual perception in the architecture. They have revolutionized image and video understanding in applications like social media, medicine, autonomous vehicles, and more.

See also:

• Deep learning
• Artificial neural network
• Feature detection
• Object recognition
• Graph neural network
• Recurrent neural network
• Attention mechanism transformer