Watershed (image processing)

In the study of image processing, a watershed is a transformation defined on a grayscale image. The name
refers metaphorically to a geological watershed, or drainage divide, which separates adjacent drainage
basins. The watershed transformation treats the image it operates upon like a topographic map, with the
brightness of each point representing its height, and finds the lines that run along the tops of ridges.

There are different technical definitions of a watershed. In graphs, watershed lines may be defined on the
nodes, on the edges, or hybrid lines on both nodes and edges. Watersheds may also be defined in the
continuous domain.[1] There are also many different algorithms to compute watersheds. Watershed
algorithms are used in image processing primarily for object segmentation purposes, that is, for separating
different objects in an image. This allows for counting the objects or for further analysis of the separated
objects.

Relief of the gradient Gradient magnitude Watershed of the Watershed of the

magnitude image gradient gradient (relief)

Definitions
In geology, a watershed is a divide that separates adjacent catchment basins.

Watershed by flooding

The idea was introduced in 1979 by S. Beucher and C. Lantuéjoul.[2] The basic idea consisted of placing a
water source in each regional minimum in the relief, to flood the entire relief from sources, and build
barriers when different water sources meet. The resulting set of barriers constitutes a watershed by flooding.
A number of improvements, collectively called Priority-Flood, have since been made to this algorithm.[3]

Watershed by topographic distance

Intuitively, a drop of water falling on a topographic relief flows towards the "nearest" minimum. The
"nearest" minimum is that minimum which lies at the end of the path of steepest descent. In terms of
topography, this occurs if the point lies in the catchment basin of that minimum. The previous definition
does not verify this condition.

Watershed by the drop of water principle

Intuitively, the watershed is a separation of the regional minima from which a drop of water can flow down
towards distinct minima. A formalization of this intuitive idea was provided in [4] for defining a watershed
of an edge-weighted graph.

Inter-pixel watershed

S. Beucher and F. Meyer introduced an algorithmic inter-pixel implementation of the watershed method,[5]
given the following procedure:

1. Label each minimum with a distinct label. Initialize a set S with the labeled nodes.
2. Extract from S a node x of minimal altitude F, that is to say F(x) = min{F(y)|y ∈ S}. Attribute
the label of x to each non-labeled node y adjacent to x, and insert y in S.
3. Repeat Step 2 until S is empty.

Topological watershed

Previous notions focus on catchment basins, but not to the produced separating line. The topological
watershed was introduced by M. Couprie and G. Bertrand in 1997,[6] and beneficiate of the following
fundamental property. A function W is a watershed of a function F if and only if W ≤ F and W preserves the
contrast between the regional minima of F; where the contrast between two regional minima M1 and M2 is
defined as the minimal altitude to which one must climb in order to go from M1 to M2 .[7] An efficient
algorithm is detailed in the paper.[8]

Watershed algorithm

Different approaches may be employed to use the watershed principle for image segmentation.

Local minima of the gradient of the image may be chosen as markers, in this case an over-
segmentation is produced and a second step involves region merging.
Marker based watershed transformation make use of specific marker positions which have
been either explicitly defined by the user or determined automatically with morphological
operators or other ways.

Meyer's flooding algorithm

One of the most common watershed algorithms was introduced by F. Meyer in the early 1990s, though a
number of improvements, collectively called Priority-Flood, have since been made to this algorithm,[9]
including variants suitable for datasets consisting of trillions of pixels.[10]
The algorithm works on a gray scale image. During the successive flooding of the grey value relief,
watersheds with adjacent catchment basins are constructed. This flooding process is performed on the
gradient image, i.e. the basins should emerge along the edges. Normally this will lead to an over-
segmentation of the image, especially for noisy image material, e.g. medical CT data. Either the image must
be pre-processed or the regions must be merged on the basis of a similarity criterion afterwards.

1. A set of markers, pixels where the flooding shall start, are chosen. Each is given a different
label.
2. The neighboring pixels of each marked area are inserted into a priority queue with a priority
level corresponding to the gradient magnitude of the pixel.
3. The pixel with the highest priority level is extracted from the priority queue. If the neighbors
of the extracted pixel that have already been labeled all have the same label, then the pixel
is labeled with their label. All non-marked neighbors that are not yet in the priority queue are
put into the priority queue.
4. Redo step 3 until the priority queue is empty.

The non-labeled pixels are the watershed lines.

Optimal spanning forest algorithms

(watershed cuts)

Watersheds as optimal spanning forest have been introduced by

Jean Cousty et al.[12] They establish the consistency of these
watersheds: they can be equivalently defined by their “catchment
basins” (through a steepest descent property) or by the “dividing
lines” separating these catchment basins (through the drop of water
principle). Then they prove, through an equivalence theorem, their
optimality in terms of minimum spanning forests. Afterward, they
introduce a linear-time algorithm to compute them. It is worthwhile
to note that similar properties are not verified in other frameworks
and the proposed algorithm is the most efficient existing algorithm,
both in theory and practice.

Example of a marker-supported
watershed transformation for a
population of pharmaceutical pellets.
Watershed lines are superimposed in
black on the CT image stack.[11]
 An image with two Result of the
markers (green), and segmentation by
a Minimum Minimum Spanning
Spanning Forest Forest
computed on the
gradient of the
image.

Links with other algorithms in computer vision

Graph cuts

In 2007, C. Allène et al.[13] established links relating Graph Cuts to optimal spanning forests. More
precisely, they show that when the power of the weights of the graph is above a certain number, the cut
minimizing the graph cuts energy is a cut by maximum spanning forest.

Shortest-path forests

The image foresting transform (IFT) of Falcao et al.[14] is a procedure for computing shortest path forests. It
has been proved by J. Cousty et al.[15] that when the markers of the IFT corresponds to extrema of the
weight function, the cut induced by the forest is a watershed cut.

Random walker

The random walker algorithm is a segmentation algorithm solving the combinatorial Dirichlet problem,
adapted to image segmentation by L. Grady in 2006.[16] In 2011, C. Couprie et al. proved that when the
power of the weights of the graph converge toward infinity, the cut minimizing the random walker energy
is a cut by maximum spanning forest.[17]

Hierarchies
A hierarchical watershed transformation converts the result into a graph display (i.e. the neighbor
relationships of the segmented regions are determined) and applies further watershed transformations
recursively. See [18] for more details. A theory linking watershed to hierarchical segmentations has been
developed in[19]

Notes
1. L. Najman and M. Schmitt. Watershed of a continuous function (https://hal-upec-upem.archiv
es-ouvertes.fr/hal-00622129/file/lpe.pdf). In Signal Processing (Special issue on
Mathematical Morphology.), Vol. 38 (1994), pages 99–112
2. Serge Beucher and Christian Lantuéj workshop on image processing, real-time edge and
motion detection (1979). http://cmm.ensmp.fr/~beucher/publi/watershed.pdf
3. Barnes, R., Lehman, C., Mulla, D., 2014. Priority-flood: An optimal depression-filling and
watershed-labeling algorithm for digital elevation models (https://arxiv.org/abs/1511.04463).
Computers & Geosciences 62, 117–127. doi:10.1016/j.cageo.2013.04.024 (https://doi.org/1
0.1016%2Fj.cageo.2013.04.024)
4. J. Cousty, G. Bertrand, L. Najman and M. Couprie. Watershed Cuts: Minimum Spanning
Forests and the Drop of Water Principle (https://hal.inria.fr/hal-01113462/document), IEEE
Transactions on Pattern Analysis and Machine Intelligence 31(8) pp. 1362-1374, 2009,
5. Serge Beucher and Fernand Meyer. The morphological approach to segmentation: the
watershed transformation (https://www.researchgate.net/profile/Serge_Beucher/publication/2
33950923_Segmentation_The_Watershed_Transformation_Mathematical_Morphology_in_I
mage_Processing/links/55f7c6ce08aeba1d9efe4072/Segmentation-The-Watershed-Transfo
rmation-Mathematical-Morphology-in-Image-Processing.pdf). In Mathematical Morphology in
Image Processing (Ed. E. R. Dougherty), pages 433–481 (1993).
6. M. Couprie, G. Bertrand. Topological gray-scale watershed transform. In Proc. of SPIE
Vision Geometry V, volume 3168, pages 136–146 (1997).
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.3.7654&rep=rep1&type=pdf
7. G. Bertrand. On topological watersheds (https://hal-upec-upem.archives-ouvertes.fr/docs/00/
62/23/98/PDF/hal.pdf). Journal of Mathematical Imaging and Vision, 22(2–3), pages 217–
230 (2005).
8. Michel Couprie, Laurent Najman, Gilles Bertrand. Quasi-linear algorithms for the topological
watershed (https://hal-upec-upem.archives-ouvertes.fr/hal-00622399/file/hal.pdf). Journal of
Mathematical Imaging and Vision, Springer Verlag, 2005, 22 (2-3), pp.231-249.
9. Barnes, R., Lehman, C., Mulla, D., 2014. Priority-flood: An optimal depression-filling and
watershed-labeling algorithm for digital elevation models (https://arxiv.org/abs/1511.04463).
Computers & Geosciences 62, 117–127. doi:10.1016/j.cageo.2013.04.024 (https://doi.org/1
0.1016%2Fj.cageo.2013.04.024)
10. Barnes, R., 2016. Parallel priority-flood depression filling for trillion cell digital elevation
models on desktops or clusters. Computers & Geosciences.
doi:10.1016/j.cageo.2016.07.001 (https://doi.org/10.1016%2Fj.cageo.2016.07.001)
11. Doerr, F. J. S., & Florence, A. J. (2020). A micro-XRT Image Analysis and Machine Learning
Methodology for the Characterisation of Multi-Particulate Capsule Formulations.
International Journal of Pharmaceutics: X, 2, 100041.
https://doi.org/10.1016/j.ijpx.2020.100041
12. Jean Cousty, Gilles Bertrand, Laurent Najman, and Michel Couprie. Watershed Cuts:
Minimum Spanning Forests and the Drop of Water Principle (https://hal.inria.fr/hal-0111346
2/document). IEEE Transactions on Pattern Analysis and Machine Intelligence. 31 (8).
August 2009. pp. 1362–1374.
13. Cédric Allène, Jean-Yves Audibert, Michel Couprie and Renaud Keriven : "Some links
between min-cuts, optimal spanning forests and watersheds (https://web.archive.org/web/20
190217100456/http://pdfs.semanticscholar.org/049b/6ea4c3e70c7249fd4538dec8dc8f24bd
de8b.pdf)", Image and Vision Computing, 2009.
14. Falcao, A.X. Stolfi, J. de Alencar Lotufo, R. : "The image foresting transform: theory,
algorithms, and applications (https://www.researchgate.net/profile/Roberto_Lotufo/publicatio
n/8331682_The_image_foresting_transform_theory_algorithms_and_applications/links/091
2f50b724229bf0d000000.pdf)", In PAMI, 2004
15. Jean Cousty, Gilles Bertrand, Laurent Najman, and Michel Couprie. Watershed cuts:
thinnings, shortest-path forests and topological watersheds (https://hal-upec-upem.archives-
ouvertes.fr/hal-00729346/document). IEEE Transactions on Pattern Analysis and Machine
Intelligence. 32 (5). 2010. pp. 925–939.
16. Grady, L.: "Random walks for image segmentation (http://leogrady.net/wp-content/uploads/2
017/01/grady2006random.pdf)". PAMI, 2006
17. Camille Couprie, Leo Grady, Laurent Najman and Hugues Talbot, "Power Watersheds: A
Unifying Graph-Based Optimization Framework (http://leogrady.net/wp-content/uploads/201
7/01/couprie2011power.pdf)”, IEEE Transactions on Pattern Analysis and Machine
Intelligence, Vol. 33, No. 7, pp. 1384-1399, July 2011
18. Laurent Najman, Michel Schmitt. Geodesic Saliency of Watershed Contours and
Hierarchical Segmentation (https://hal-upec-upem.archives-ouvertes.fr/hal-00622128/docum
ent). IEEE Transactions on Pattern Analysis and Machine Intelligence, Institute of Electrical
and Electronics Engineers, 1996, 18 (12), pp.1163-1173.
19. Laurent Najman. On the equivalence between hierarchical segmentations and ultrametric
watersheds (https://arxiv.org/abs/1002.1887). Journal of Mathematical Imaging and Vision,
Springer Verlag, 2011, 40 (3), pp.231-247.

References
Fernand Meyer. Un algorithme optimal pour la ligne de partage des eaux. Dans 8me congrès
de reconnaissance des formes et intelligence artificielle, Vol. 2 (1991), pages 847–857,
Lyon, France.
Luc Vincent and Pierre Soille. Watersheds in digital spaces: an efficient algorithm based on
immersion simulations (https://web.archive.org/web/20190802193957/https://pdfs.semantics
cholar.org/a381/9dda9a5f00dbb8cd3413ca7422e37a0d5794.pdf). In IEEE Transactions on
Pattern Analysis and Machine Intelligence, Vol. 13, Num. 6 (1991), pages 583–598.
L. Najman and M. Schmitt. Geodesic saliency of watershed contours and hierarchical
segmentation (https://hal-upec-upem.archives-ouvertes.fr/hal-00622128/document). In IEEE
Transactions on Pattern Analysis and Machine Intelligence, Vol. 18, Num. 12 (1996), pages
1163–1173.
J.B.T.M. Roerdink and A. Meijster. The watershed transform: definitions, algorithms, and
parallelization strategies (http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.10.3852
&rep=rep1&type=pdf). In Fundamenta Informaticae 41 (2000), pp. 187–228.
Laurent Najman, Michel Couprie and Gilles Bertrand. Watersheds, mosaics, and the
emergence paradigm (https://www.sciencedirect.com/science/article/pii/S0166218X040036
6X). In Discrete Applied Mathematics, Vol. 147, Num. 2–3(2005), Pages 301–324.

External links
The Watershed Transformation (http://cmm.ensmp.fr/~beucher/wtshed.html) with animations
of the watershed algorithm.
 Topological Watershed Transform (http://perso.esiee.fr/~coupriem/tw/) with papers, lecture
slides and source code.
An open source watershed plugin (http://rsbweb.nih.gov/ij/plugins/watershed.html) for
ImageJ.
The Topology ToolKit (https://topology-tool-kit.github.io/) (2D and 3D watersheds based on
the Morse complex)

Retrieved from "https://en.wikipedia.org/w/index.php?title=Watershed_(image_processing)&oldid=1162344268"