Learned Watershed: End-to-End Learning of Seeded Segmentation

@article{Wolf2017LearnedWE,
  title={Learned Watershed: End-to-End Learning of Seeded Segmentation},
  author={Steffen Wolf and Lukas Schott and U. K{\"o}the and Fred A. Hamprecht},
  journal={2017 IEEE International Conference on Computer Vision (ICCV)},
  year={2017},
  pages={2030-2038}
}
Learned boundary maps are known to outperform handcrafted ones as a basis for the watershed algorithm. [] Key Method The estimator for the merging priorities is cast as a neural network that is convolutional (over space) and recurrent (over iterations). The latter allows learning of complex shape priors. The method gives the best known seeded segmentation results on the CREMI segmentation challenge.
View on IEEE
View With Semantic Reader
28 Citations
Background Citations
7
Methods Citations
8
Results Citations
2

Figures and Tables from this paper

28 Citations

End-To-End Learned Random Walker for Seeded Image Segmentation
TLDR
This work presents an end-to-end learned algorithm based on the Random Walker algorithm, where the edge weights of the un- derlying graph are predicted using a convolutional neural network to predict context-dependent diffusivities for a linear diffusion process.
Probabilistic Deep Learning for Instance Segmentation
TLDR
This work proposes a generic method to obtain model-inherent uncertainty estimates within proposal-free instance segmentation models and analyzes the quality of the uncertainty estimates with a metric adapted from semantic segmentation.
Large Scale Image Segmentation with Structured Loss Based Deep Learning for Connectome Reconstruction
TLDR
A method combining affinity prediction with region agglomeration, which improves significantly upon the state of the art of neuron segmentation from electron microscopy (EM) in accuracy and scalability, and suggests that a single method can be applied to both nearly isotropic block-face EM data and anisotropic serial sectioned EM data.
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.
CNN-Based Preprocessing to Optimize Watershed-Based Cell Segmentation in 3D Confocal Microscopy Images
TLDR
A new 3D segmentation approach is proposed that combines the discriminative power of convolutional neural networks for preprocessing and investigates the performance of three watershed-based postprocessing strategies, which are well suited to segment object shapes, even when supplied with vague seed and boundary constraints.
Deep morphological networks
PatchPerPix for Instance Segmentation
TLDR
A novel method for proposal free instance segmentation that can handle sophisticated object shapes which span large parts of an image and form dense object clusters with crossovers, and is the first non-iterative method that yields instances that are composed of learnt shape patches.
Probabilistic Watershed: Sampling all spanning forests for seeded segmentation and semi-supervised learning
TLDR
A simpler proof of this equivalence and establish the computational feasibility of the Probabilistic Watershed with Kirchhoff's matrix tree theorem are given and a new connection between the Random Walker probabilities and the triangle inequality of the effective resistance is shown.
The Mutex Watershed and its Objective: Efficient, Parameter-Free Graph Partitioning
TLDR
The Mutex Watershed is proposed, an efficient algorithm for graph partitioning that can accommodate not only attractive but also repulsive cues, allowing it to find a previously unspecified number of segments without the need for explicit seeds or a tunable threshold.
Watersheds for Semi-Supervised Classification
TLDR
A framework is developed, which allows one to answer questions about the boundaries of the watershed operator and shows that watersheds form a particular class of morphMedian classifiers, resulting in the maximum margin principle.
...
...

References

SHOWING 1-10 OF 35 REFERENCES
Deep Watershed Transform for Instance Segmentation
  • Min Bai, R. Urtasun
  • Computer Science
    2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2017
TLDR
This paper presents a simple yet powerful end-to-end convolutional neural network that achieves more than double the performance over the state-of-the-art on the challenging Cityscapes Instance Level Segmentation task.
Flood-Filling Networks
TLDR
A unified and end-to-end trainable machine learning approach, flood-filling networks, in which a recurrent 3d convolutional network directly produces individual segments from a raw image, which robustly segments images with an unknown and variable number of objects as well as highly variable object sizes.
U-Net: Convolutional Networks for Biomedical Image Segmentation
TLDR
It is shown that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks.
Pushing the Boundaries of Boundary Detection using Deep Learning
TLDR
This work shows that adapting Deep Convolutional Neural Network training to the task of boundary detection can result in substantial improvements over the current state-of-the-art in boundary detection, and examines the potential of the boundary detector in conjunction with thetask of semantic segmentation.
Machine Learning of Hierarchical Clustering to Segment 2D and 3D Images
TLDR
The use of variation of information to measure segmentation accuracy is advocated, particularly in 3D electron microscopy images of neural tissue, and using this metric demonstrate an improvement over competing algorithms in EM and natural images.
Crowdsourcing the creation of image segmentation algorithms for connectomics
To stimulate progress in automating the reconstruction of neural circuits, we organized the first international challenge on 2D segmentation of electron microscopic (EM) images of the brain.…
Convolutional Networks Can Learn to Generate Affinity Graphs for Image Segmentation
TLDR
This work presents a machine learning approach to computing an affinity graph using a convolutional network (CN) trained using ground truth provided by human experts and shows that the CN affinity graph can be paired with any standard partitioning algorithm and improves segmentation accuracy significantly compared to standard hand-designed affinity functions.
Holistically-Nested Edge Detection
  • Saining Xie, Z. Tu
  • Computer Science
    2015 IEEE International Conference on Computer Vision (ICCV)
  • 2015
TLDR
HED performs image-to-image prediction by means of a deep learning model that leverages fully convolutional neural networks and deeply-supervised nets, and automatically learns rich hierarchical representations that are important in order to resolve the challenging ambiguity in edge and object boundary detection.
Image Segmentation by Size-Dependent Single Linkage Clustering of a Watershed Basin Graph
We present a method for hierarchical image segmentation that defines a disaffinity graph on the image, over-segments it into watershed basins, defines a new graph on the basins, and then merges…
Stochastic watershed segmentation
This paper introduces a watershed-based stochastic segmentation methodology. The approach is based on using M realizations of N random markers to build a probability density function (pdf) of…
...
...