NAME
Image::Leptonica::Func::watershed
VERSION
version 0.04
watershed.c
watershed.c
Top-level
L_WSHED *wshedCreate()
void wshedDestroy()
l_int32 wshedApply()
Helpers
static l_int32 identifyWatershedBasin()
static l_int32 mergeLookup()
static l_int32 wshedGetHeight()
static void pushNewPixel()
static void popNewPixel()
static void pushWSPixel()
static void popWSPixel()
static void debugPrintLUT()
static void debugWshedMerge()
Output
l_int32 wshedBasins()
PIX *wshedRenderFill()
PIX *wshedRenderColors()
The watershed function identifies the "catch basins" of the input
8 bpp image, with respect to the specified seeds or "markers".
The use is in segmentation, but the selection of the markers is
critical to getting meaningful results.
How are the markers selected? You can't simply use the local
minima, because a typical image has sufficient noise so that
a useful catch basin can easily have multiple local minima. However
they are selected, the question for the watershed function is
how to handle local minima that are not markers. The reason
this is important is because of the algorithm used to find the
watersheds, which is roughly like this:
(1) Identify the markers and the local minima, and enter them
into a priority queue based on the pixel value. Each marker
is shrunk to a single pixel, if necessary, before the
operation starts.
(2) Feed the priority queue with neighbors of pixels that are
popped off the queue. Each of these queue pixels is labelled
with the index value of its parent.
(3) Each pixel is also labelled, in a 32-bit image, with the marker
or local minimum index, from which it was originally derived.
(4) There are actually 3 classes of labels: seeds, minima, and
fillers. The fillers are labels of regions that have already
been identified as watersheds and are continuing to fill, for
the purpose of finding higher watersheds.
(5) When a pixel is popped that has already been labelled in the
32-bit image and that label differs from the label of its
parent (stored in the queue pixel), a boundary has been crossed.
There are several cases:
(a) Both parents are derived from markers but at least one
is not deep enough to become a watershed. Absorb the
shallower basin into the deeper one, fixing the LUT to
redirect the shallower index to the deeper one.
(b) Both parents are derived from markers and both are deep
enough. Identify and save the watershed for each marker.
(c) One parent was derived from a marker and the other from
a minima: absorb the minima basin into the marker basin.
(d) One parent was derived from a marker and the other is
a filler: identify and save the watershed for the marker.
(e) Both parents are derived from minima: merge them.
(f) One parent is a filler and the other is derived from a
minima: merge the minima into the filler.
(6) The output of the watershed operation consists of:
- a pixa of the basins
- a pta of the markers
- a numa of the watershed levels
Typical usage:
L_WShed *wshed = wshedCreate(pixs, pixseed, mindepth, 0);
wshedApply(wshed);
wshedBasins(wshed, &pixa, &nalevels);
... do something with pixa, nalevels ...
pixaDestroy(&pixa);
numaDestroy(&nalevels);
Pix *pixd = wshedRenderFill(wshed);
wshedDestroy(&wshed);
FUNCTIONS
wshedApply
l_int32 wshedApply ( L_WSHED *wshed )
wshedApply()
Input: wshed (generated from wshedCreate())
Return: 0 if OK, 1 on error
Iportant note:
(1) This is buggy. It seems to locate watersheds that are
duplicates. The watershed extraction after complete fill
grabs some regions belonging to existing watersheds.
See prog/watershedtest.c for testing.
wshedBasins
l_int32 wshedBasins ( L_WSHED *wshed, PIXA **ppixa, NUMA **pnalevels )
wshedBasins()
Input: wshed
&pixa (<optional return> mask of watershed basins)
&nalevels (<optional return> watershed levels)
Return: 0 if OK, 1 on error
wshedCreate
L_WSHED * wshedCreate ( PIX *pixs, PIX *pixm, l_int32 mindepth, l_int32 debugflag )
wshedCreate()
Input: pixs (8 bpp source)
pixm (1 bpp 'marker' seed)
mindepth (minimum depth; anything less is not saved)
debugflag (1 for debug output)
Return: WShed, or null on error
Notes:
(1) It is not necessary for the fg pixels in the seed image
be at minima, or that they be isolated. We extract a
single pixel from each connected component, and a seed
anywhere in a watershed will eventually label the watershed
when the filling level reaches it.
(2) Set mindepth to some value to ignore noise in pixs that
can create small local minima. Any watershed shallower
than mindepth, even if it has a seed, will not be saved;
It will either be incorporated in another watershed or
eliminated.
wshedDestroy
void wshedDestroy ( L_WSHED **pwshed )
wshedDestroy()
Input: &wshed (<will be set to null before returning>)
Return: void
wshedRenderColors
PIX * wshedRenderColors ( L_WSHED *wshed )
wshedRenderColors()
Input: wshed
Return: pixd (initial image with all basins filled), or null on error
wshedRenderFill
PIX * wshedRenderFill ( L_WSHED *wshed )
wshedRenderFill()
Input: wshed
Return: pixd (initial image with all basins filled), or null on error
AUTHOR
Zakariyya Mughal <zmughal@cpan.org>
COPYRIGHT AND LICENSE
This software is copyright (c) 2014 by Zakariyya Mughal.
This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.