diff options
author | Nick White <git@njw.name> | 2020-08-03 17:20:58 +0100 |
---|---|---|
committer | Nick White <git@njw.name> | 2020-08-03 17:20:58 +0100 |
commit | 30b54ffce25c86f3e1c58076f39ef9ace53b0274 (patch) | |
tree | 83b8ebf7a5bb516f12ff55a2c102137933aecdfc | |
parent | 26a5872f3470f00d438b4f9d720b712ef55303c1 (diff) |
Improve documentationv0.5.0
-rw-r--r-- | integralimg.go | 13 |
1 files changed, 10 insertions, 3 deletions
diff --git a/integralimg.go b/integralimg.go index 6cf3a25..d4d25f1 100644 --- a/integralimg.go +++ b/integralimg.go @@ -7,9 +7,16 @@ // sum of pixels to the left and above each pixel, which can make // several common image processing operations much faster. // -// The Sum(), Mean() and MeanStdDev() functions provided for the -// integral versions of Images significantly speed up many common -// image processing operations. +// A lot of image processing operations rely on many calculations +// of the sum or mean of a set of pixels. As these have been +// precalculated for an integral Image, these calculations are +// much faster. Image.Sum() and Image.Mean() functions are provided +// by this package to take advantage of this. +// +// Another common requirement is standard deviation over an area +// of an image. This can be calculated by creating an integral +// Image and squared integral Image (SqImage) for a base image, and +// passing them to the MeanStdDev() function provided. package integralimg import ( |