Anisotropic EM Segmentation by 3D Affinity Learning and Agglomeration

Parag, Toufiq; Tschopp, Fabian; Grisaitis, William; Turaga, Srinivas C.; Zhang, Xuewen; Matejek, Brian; Kamentsky, Lee; Lichtman, Jeff W.; Pfister, Hanspeter

doi:10.48550/arxiv.1707.08935

Cited by 5 publications

(9 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, the dense gradient of the encoded-neighborhood branch, which focuses on locally correct predictions, nicely complements the sparse gradient 5 of the encoded-neighborhood branch, which focuses on predictions that are consistent in a larger neighborhood. We expect this [34] 3D-Watershed 0.566 LFC [22] Z-Watershed+Agglo 0.616…”

Section: Resultsmentioning

confidence: 98%

See 1 more Smart Citation

Proposal-Free Volumetric Instance Segmentation from Latent Single-Instance Masks

Bailoni¹,

Pape²,

Wolf³

et al. 2020

Preprint

View full text Add to dashboard Cite

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. In contrast to related approaches, our method concurrently predicts all masks, one for each pixel, and thus resolves any conflict jointly across the entire image. Specifically, predictions from overlapping masks are combined into edge weights of a signed graph that is subsequently partitioned to obtain all final instances concurrently. The result is a parameter-free method that is strongly robust to noise and prioritizes predictions with the highest consensus across overlapping masks. All masks are decoded from a low dimensional latent representation, which results in great memory savings strictly required for applications to large volumetric images. We test our method on the challenging CREMI 2016 neuron segmentation benchmark where it achieves competitive scores.

show abstract

Section: Resultsmentioning

confidence: 98%

“…Predicting Pixel-Pair Affinities -Instance-aware edge detection has experienced recent progress due to deep learning, both on natural images and biological data [8,18,15,32,28,34,22,3]. Among these methods, the most recent ones also predict long-range affinities between pixels and not only direct-neighbor relationships [8,18,15].…”

Section: Related Workmentioning

confidence: 99%

Proposal-Free Volumetric Instance Segmentation from Latent Single-Instance Masks

Bailoni¹,

Pape²,

Wolf³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…[81] 0.322 GASP Max. + Constraints 0.324 GASP Sum [36] 0.334 GASP Average + Constraints 0.563 THRESH 1.521 [26] 0.276 CRU-Net [85] 0.566 LFC [64] 0.616 predict graph edge weights, relying not only on information derived from affinity maps but also raw data and shape information. Note that the test volumes contain several imaging artifacts that make segmentation particularly challenging and might profit from more robust edge statistics of super-pixel based approaches.…”

Section: Resultsmentioning

confidence: 99%

A Generalized Framework for Agglomerative Clustering of Signed Graphs applied to Instance Segmentation

Bailoni¹,

Pape²,

Hütsch³

et al. 2019

Preprint

View full text Add to dashboard Cite

We propose a novel theoretical framework that generalizes algorithms for hierarchical agglomerative clustering to weighted graphs with both attractive and repulsive interactions between the nodes. This framework defines GASP, a Generalized Algorithm for Signed graph Partitioning, and allows us to explore many combinations of different linkage criteria and cannot-link constraints. We prove the equivalence of existing clustering methods to some of those combinations, and introduce new algorithms for combinations which have not been studied. An extensive comparison is performed to evaluate properties of the clustering algorithms in the context of instance segmentation in images, including robustness to noise and efficiency. We show how one of the new algorithms proposed in our framework outperforms all previously known agglomerative methods for signed graphs, both on the competitive CREMI 2016 EM segmentation benchmark and on the CityScapes dataset.Preprint. Under review.

show abstract

“…Our watershed code was adopted from [35]. Similar to other traditional agglomerating-techniques, Neuroproof trains a random forest on merge decisions of neighboring objects [40,41,42]. These baseline methods were fed with the same high-quality border maps used in our 3C reconstruction system.…”

Section: Snemi3d Datasetmentioning

confidence: 99%

“…For the second test, we were granted permission to reconstruct a recently collected rat cortex dataset of the Lichtman group at Harvard (ECS). This test allowed the comparison of 3C to the excellent agglomerative approach of [42] (4th on SNEMI3D), while using exactly their U-Net [50] border predictions as inputs to our 3C network. On the test set our NRI score was 0.86, compared to 0.73 for the agglomeration pipeline.…”

Section: Harvard Rodent Cortex Datasets (Ecs S1)mentioning

confidence: 99%

Cross-Classification Clustering: An Efficient Multi-Object Tracking Technique for 3-D Instance Segmentation in Connectomics

Meirovitch

Saribekyan

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

Pixel-accurate tracking of objects is a key element in many computer vision applications, often solved by iterated individual object tracking or instance segmentation followed by object matching. Here we introduce crossclassification clustering (3C), a technique that simultaneously tracks complex, interrelated objects in an image stack.The key idea in cross-classification is to efficiently turn a clustering problem into a classification problem by running a logarithmic number of independent classifications per image, letting the cross-labeling of these classifications uniquely classify each pixel to the object labels. We apply the 3C mechanism to achieve state-of-the-art accuracy in connectomics -the nanoscale mapping of neural tissue from electron microscopy volumes. Our reconstruction system increases scalability by an order of magnitude over existing single-object tracking methods (such as flood-filling networks). This scalability is important for the deployment of connectomics pipelines, since currently the best performing techniques require computing infrastructures that are beyond the reach of most laboratories. Our algorithm may offer benefits in other domains that require pixel-accurate tracking of multiple objects, such as segmentation of videos and medical imagery. * These authors equally contributed to this work

show abstract

Anisotropic EM Segmentation by 3D Affinity Learning and Agglomeration

Cited by 5 publications

References 0 publications

Proposal-Free Volumetric Instance Segmentation from Latent Single-Instance Masks

Proposal-Free Volumetric Instance Segmentation from Latent Single-Instance Masks

A Generalized Framework for Agglomerative Clustering of Signed Graphs applied to Instance Segmentation

Cross-Classification Clustering: An Efficient Multi-Object Tracking Technique for 3-D Instance Segmentation in Connectomics

Contact Info

Product

Resources

About