by Michele Laurelli
A small matrix of learnable weights that slides over input during convolution to detect specific features.
Filters are learned during training to detect useful features. Common sizes: 3x3, 5x5, 7x7. Multiple filters extract different features. The term 'kernel' is used interchangeably with filter.
3x3 edge detection filter
5x5 Gaussian blur kernel
Learned feature detectors
A mathematical operation that slides a filter/kernel over input data to extract features.
A deep learning architecture specialized for processing grid-like data such as images, using convolutional layers.
The output of applying a convolution filter to an input, representing detected features.