Convolutional Neural Network (CNN)
A neural network architecture that uses learnable filters to detect spatial patterns in data, primarily used for image processing and computer vision tasks.
How It Works
CNNs apply small learnable filters (kernels) that slide across the input image, detecting local features like edges, textures, and shapes. Deeper layers combine these features into increasingly complex patterns (eyes, faces, objects).
Key Components
Convolutional layers: Apply filters to detect features. Pooling layers: Downsample to reduce size and add translation invariance. Fully connected layers: Final classification.
Landmark Architectures
LeNet (1998), AlexNet (2012, sparked deep learning revolution), VGG (2014), ResNet (2015, introduced skip connections), EfficientNet (2019). Vision Transformers (ViT) now compete with or surpass CNNs on many tasks.