In simple words: LeNet made CNNs practical for digit recognition.
What LeNet did
LeNet is an early Convolutional Neural Network known for handwritten digit recognition. It used repeated convolution and pooling followed by layers that classified the digit.
Why it matters
It showed that a network could learn useful image features from examples instead of relying only on hand-designed rules. It was built for simpler, smaller images than many modern vision tasks.
See the flow
digit image→convolution→pooling→classification
Simple example
A scanned cheque contains a handwritten “5”. LeNet-style processing finds strokes and curves, then predicts the digit.
Remember: LeNet made CNNs practical for digit recognition.