Convolution kernel

Applies to: general, python

A convolution kernel is a small set of weights slid over data to filter, blur, detect edges, or compute local features.

y[i] = np.sum(x[i:i+3] * kernel)

See also: convolution, fft