Convolutional Neural Network (CNN)
Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary
A neural network built for grid-shaped data such as images, using convolutional filters that slide across the input to learn spatial patterns layer by layer.
What is a Convolutional Neural Network (CNN)?
A Convolutional Neural Network (CNN) is a deep learning model built for grid-shaped data such as images, using filters that slide across the input to detect local patterns like edges and textures.
The trick is weight sharing: instead of learning a separate parameter for every pixel, a CNN reuses one small filter across the whole image. A cat detector learned in the top-left corner works in the bottom-right too, which keeps the network small enough to train on real photos.
How a CNN Works
- Convolutional layers slide filters over the input and produce feature maps. Early layers respond to edges and color blobs; deeper layers combine those into shapes, parts, and objects.
- Pooling layers shrink each feature map, keeping the strongest signals while cutting computation and making the network less sensitive to small shifts in position.
- Fully connected layers sit at the end and turn the extracted features into a final prediction, such as a class label.
CNN vs RNN
The practical difference: a CNN reads spatial data all at once, while a recurrent neural network reads sequential data one step at a time. A CNN asks “which pixels sit next to which”; an RNN asks “what came before this”.
| Criterion | CNN | RNN |
|---|---|---|
| Data shape | Grid-like: images, video frames, spectrograms | Sequential: text, audio, time series |
| What it exploits | Spatial locality between neighboring values | Order and context from earlier steps |
| Processing | Whole input in parallel | Step by step, carrying a hidden state |
| Typical tasks | Image classification, object detection | Language modeling, forecasting |
Applications of CNNs
- Image recognition: Assigning a label to a whole image, such as separating cats from dogs.
- Object detection: Locating and identifying multiple objects in one scene, the core of a self-driving car’s vision stack.
- Medical imaging: Flagging abnormalities in X-rays, MRIs, and CT scans for a radiologist to review.
Example of a Convolutional Neural Network
AlexNet, a CNN trained by Alex Krizhevsky and colleagues, won the 2012 ImageNet image-classification competition by a wide margin over non-neural methods. That result convinced much of the field that deep learning was the path forward for computer vision.
Follow a photo through it: convolutional filters first fire on edges and textures, deeper layers assemble those into whiskers and ears, pooling discards position noise along the way, and the final layers output “cat” with a confidence score.
Related AI terms: Recurrent Neural Network · Model Architecture · Deep Learning · Computer Vision · Image Recognition
Did you like the Convolutional Neural Network (CNN) gist?
Learn about 250+ need-to-know artificial intelligence terms in the AI Dictionary.
Mihail Sebastian — Writes about AI governance, regulation, and the technology behind them. Placeholder bio — replace with a real credential line. About