The Mutex Watershed and its Objective: Efficient, Parameter-Free Graph Partitioning

@article{Wolf2021TheMW,
  title={The Mutex Watershed and its Objective: Efficient, Parameter-Free Graph Partitioning},
  author={Steffen Wolf and Alberto Bailoni and Constantin Pape and Nasim Rahaman and Anna Kreshuk and U. K{\"o}the and Fred A. Hamprecht},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year={2021},
  volume={43},
  pages={3724-3738}
}
Image partitioning, or segmentation without semantics, is the task of decomposing an image into distinct segments, or equivalently to detect closed contours. Most prior work either requires seeds, one per segment; or a threshold; or formulates the task as multicut / correlation clustering, an NP-hard problem. Here, we propose an efficient algorithm for graph partitioning, the “Mutex Watershed”. Unlike seeded watershed, the algorithm can accommodate not only attractive but also repulsive cues… 

Figures and Tables from this paper

The Semantic Mutex Watershed for Efficient Bottom-Up Semantic Instance Segmentation
TLDR
A greedy algorithm for joint graph partitioning and labeling derived from the efficient Mutex Watershed partitioning algorithm that optimizes an objective function closely related to the Symmetric Multiway Cut objective and empirically shows efficient scaling behavior.
An Efficient Hierarchical Layered Graph Approach for Multi-Region Segmentation
TLDR
Compared to hierarchical segmentation by the min-cut/max-flow algorithm, the proposed HLOIFT approach is less restrictive, leading to globally optimal results in more general scenarios, and has a better running time.
Efficient Hierarchical Multi-Object Segmentation in Layered Graphs
TLDR
Compared to the hierarchical segmentation by the min-cut/max-flow algorithm, this approach is less restrictive, leading to globally optimal results in more general scenarios, and has a better running time.
GASP, a generalized framework for agglomerative clustering of signed graphs and its application to Instance Segmentation
TLDR
A theoretical framework that generalizes simple and fast algorithms for hierarchical agglomerative clustering to weighted graphs with both attractive and repulsive interactions between the nodes and achieves state-of-theart accuracy on the CREMI 2016 EM segmentation benchmark without requiring domain-specific superpixels.
Two Polynomial Time Graph Labeling Algorithms Optimizing Max-Norm-Based Objective Functions
TLDR
This work considers a problem of finding an optimal binary labeling for the objective function defined as the max-norm over a set of local costs of a form that naturally appears in image processing, and proposes two algorithms that find such optimal labelings.
RAMA: A Rapid Multicut Algorithm on GPU
TLDR
This work proposes a highly parallel primal-dual algorithm for the multicut problem, a classical graph clustering problem widely used in machine learning and computer vision, and shows resulting one to two orders ofmagnitudes improvements in execution speed without sacrificing solution quality compared to traditional sequential algorithms that run on CPUs.
A tutorial on applications of power watershed optimization to image processing
TLDR
The intuition behind the working of PW framework, i.e. exploitation of contrast invariance on image data is explained and the ideas from the article can be adapted to a large class of graph-based cost minimization methods to obtain scalable segmentation and filtering algorithms.
Proposal-Free Volumetric Instance Segmentation from Latent Single-Instance Masks
TLDR
This work introduces a new proposal-free instance segmentation method that builds on single-instance segmentation masks predicted across the entire image in a sliding window style that results in great memory savings strictly required for applications to large volumetric images.
Leveraging Domain Knowledge to improve EM image segmentation with Lifted Multicuts
TLDR
It is shown how domain-specific information can be leveraged by expressing it as long-range interactions in a graph partitioning problem known as the lifted multicut problem, demonstrating significant improvement in segmentation accuracy for three challenging EM segmentation problems from neuroscience and cell biology.
...
...

References

SHOWING 1-10 OF 84 REFERENCES
The Mutex Watershed: Efficient, Parameter-Free Image Partitioning
TLDR
When presented with short-range attractive and long-range repulsive cues from a deep neural network, the Mutex Watershed gives results that currently define the state-of-the-art in the competitive ISBI 2012 EM segmentation benchmark.
Power Watershed: A Unifying Graph-Based Optimization Framework
TLDR
This work extends a common framework for graph-based image segmentation that includes the graph cuts, random walker, and shortest path optimization algorithms and proposes a new family of segmentation algorithms that fixes p to produce an optimal spanning forest but varies the power q beyond the usual watershed algorithm.
Efficient Decomposition of Image and Mesh Graphs by Lifted Multicuts
TLDR
This work proposes a generalization of the MP with long-range terms (LMP) and designs and implements two efficient algorithms (primal feasible heuristics) for the MP and LMP which allow us to study instances of both problems w.r.t. the pixel grid graphs of the images in the BSDS-500 benchmark.
Generalized Hard Constraints for Graph Segmentation
TLDR
This work proposes a new type of hard constraints, that includes both regional constraints and boundary constraints as special cases, and presents an efficient method for computing cuts that satisfy a set of generalized constraints, while globally minimizing a graph cut measure.
Constrained connectivity for hierarchical image partitioning and simplification
  • P. Soille
  • Computer Science
    IEEE Transactions on Pattern Analysis and Machine Intelligence
  • 2008
TLDR
An image partitioning and simplification method based on the constrained connectivity paradigm that includes a generalization to multichannel images, application examples, a review of related image segmentation techniques, and pseudocode for an implementation based on queue and stack data structures are introduced.
Global-local optimizations by hierarchical cuts and climbing energies
Maximin affinity learning of image segmentation
TLDR
This work presents the first machine learning algorithm for training a classifier to produce affinity graphs that are good in the sense of producing segmentations that directly minimize the Rand index, a well known segmentation performance measure.
Watershed Cuts: Minimum Spanning Forests and the Drop of Water Principle
TLDR
The consistency of these watersheds is established, and it is proved, through an equivalence theorem, their optimality in terms of minimum spanning forests and the proposed algorithm is the most efficient existing algorithm, both in theory and in practice.
Globally Optimal Closed-Surface Segmentation for Connectomics
TLDR
A practical cutting-plane approach to solve the MAP inference problem to global optimality despite its NP-hardness is proposed and this approach is applied to challenging large-scale 3D segmentation problems for neural circuit reconstruction (Connectomics), demonstrating the advantage of this higher-order model over independent decisions and finite-order approximations.
A Multi-cut Formulation for Joint Segmentation and Tracking of Multiple Objects
TLDR
This paper proposes a joint graphical model for point trajectories and object detections whose Multicuts are solutions to motion segmentation and multi-target tracking problems at once and argues that high-level cues from object detection and tracking are helpful to solve motion segmentsation.
...
...