summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorNick White <git@njw.name>2020-11-02 11:40:22 +0000
committerNick White <git@njw.name>2020-11-02 11:40:22 +0000
commit4779d2cdaa89d06ce88d2c7de968f3f861821454 (patch)
treeea5023e0bc06476192e31211870a6ff2a8dc2902
parente5fbc00de99f2f641106ea47f413878848fa9709 (diff)
wip adaptive binarisation
-rw-r--r--content/posts/adaptive-binarisation/index.md9
1 files changed, 9 insertions, 0 deletions
diff --git a/content/posts/adaptive-binarisation/index.md b/content/posts/adaptive-binarisation/index.md
index 4266343..e4e2232 100644
--- a/content/posts/adaptive-binarisation/index.md
+++ b/content/posts/adaptive-binarisation/index.md
@@ -29,3 +29,12 @@ threshold diagram was])
and local thresholding, setting the global one such that the image is
screwed up)
-->
+
+A popular algorithm for this adaptive binarisation technique was
+described in a 2000 paper by J. Sauvola
+[Adaptive document image binarization](http://www.ee.oulu.fi/mvg/files/pdf/pdf_24.pdf),
+and is now generally just refered to as the "Sauvola algorithm".
+
+only some of what sauvola actually outlined in the paper is generally used today, which is a modification of niblack's (1986). sauvola's paper suggests several different binarisation methods, with an algorithm to switch between them, but these days it's just the main "binarization of textual components" part (3.3) that the paper is remembered for, and is generally simply refered to as "sauvola".
+
+there are two variables, 'k', the so-called 'threshold value', and the window size.