Morphological Error Detection in 3D Segmentations

05/30/2017
by   David Rolnick, et al.
0

Deep learning algorithms for connectomics rely upon localized classification, rather than overall morphology. This leads to a high incidence of erroneously merged objects. Humans, by contrast, can easily detect such errors by acquiring intuition for the correct morphology of objects. Biological neurons have complicated and variable shapes, which are challenging to learn, and merge errors take a multitude of different forms. We present an algorithm, MergeNet, that shows 3D ConvNets can, in fact, detect merge errors from high-level neuronal morphology. MergeNet follows unsupervised training and operates across datasets. We demonstrate the performance of MergeNet both on a variety of connectomics data and on a dataset created from merged MNIST images.

READ FULL TEXT

page 5

page 8

research
11/30/2020

Modelling Verbal Morphology in Nen

Nen verbal morphology is remarkably complex; a transitive verb can take ...
research
09/03/2021

MitoVis: A Visually-guided Interactive Intelligent System for Neuronal Mitochondria Analysis

Neurons have a polarized structure, including dendrites and axons, and c...
research
02/25/2022

Morphology Without Borders: Clause-Level Morphological Annotation

Morphological tasks use large multi-lingual datasets that organize words...
research
04/06/2022

Urdu Morphology, Orthography and Lexicon Extraction

Urdu is a challenging language because of, first, its Perso-Arabic scrip...
research
10/17/2018

Characterization of Brain Cortical Morphology Using Localized Topology-Encoding Graphs

The human brain cortical layer has a convoluted morphology that is uniqu...
research
03/18/2021

Cellcounter: a deep learning framework for high-fidelity spatial localization of neurons

Many neuroscientific applications require robust and accurate localizati...
research
05/20/2020

Biologically-Constrained Graphs for Global Connectomics Reconstruction

Most current state-of-the-art connectome reconstruction pipelines have t...

Please sign up or login with your details

Forgot password? Click here to reset