Convolution and cross correlation functions for ndarrays

Downloads in past


2.0.09 years ago11 years agoMinified + gzip package size for ndarray-convolve in KB


ndarray-convolve ================ Convolutions and cross correlations on ndarrays.


```javascript //Read in test image var lena = require("luminance")(require("lena")) //Generate a 5-point Laplace filter var filter = require("ndarray-pack")(0, 1, 0,
[1, -4, 1],
[0, 1, 0]])
//Convolve them together require("ndarray-convolve")(lena, filter) ``` This produces the following array:


npm install ndarray-convolve


```javascript var convolve = require("ndarray-convolve") ```

convolve( ... )

Performs a convolution between two images with zero boundary conditions. As long as it does not cause unnecessary cropping, the kernel (b) will be assumed to have its origin in the center of the kernel (for even kernels, slightly to the right of the center, for example [1,2,3,4] would be assumed to have it origin at 3). There are four ways you can call this function:

convolve(a, b)

Convolves a and b storing the result in a

convolve(out, a, b)

Convolves a and b storing the result in out

convolve(a_r, a_i, b_r, b_i)

Convolves two complex arrays storing the result in a_r, a_i

convolve(out_r, out_i, a_r, a_i, b_r, b_i)

Convolves two complex array storing the result in out_r, out_i

convolve.wrap( ... )

Convolves two arrays with periodic boundary conditions. Same convention as convolve

convolve.correlate( ... )

Cross correlates two arrays with zero boundary conditions. Same convention.

convolve.correlate.wrap( ... )

Cross correlates two arrays with wrapped boundary conditions. Same convention again.


(c) 2013 Mikola Lysenko. MIT License