What is a convolutional neural network and why is convolution well-suited to images?

A CNN is a network built around convolutional layers that slide small learnable filters (kernels) across the input, computing dot products to produce feature maps. This suits images for three reasons. First, parameter sharing: the same filter is reused across all spatial positions, drastically reducing parameters versus a fully connected layer and detecting a feature wherever it appears. Second, local connectivity: each output depends only on a small receptive field, matching the fact that image structure is local. Third, translation equivariance: shifting the input shifts the feature map, so learned patterns generalize across positions. Stacked convolutions build a hierarchy from edges to textures to objects. Pooling layers downsample to add spatial invariance and cut computation. CNNs power image classification, detection, and segmentation, and the same ideas apply to audio and time-series data.