Selective labeling: identifying representative sub-volumes for interactive segmentation

Automatic segmentation of challenging biomedical volumes with multiple objects is still an open research field. Automatic approaches usually require a large amount of training data to be able to model the complex and often noisy appearance and structure of biological organelles and their boundaries....

Full description

Bibliographic Details
Main Authors: Luengo, Imanol, Basham, Mark, French, Andrew P.
Format: Article
Published: Springer Verlag 2016
Subjects:
Online Access:https://eprints.nottingham.ac.uk/44843/
_version_ 1848797010952454144
author Luengo, Imanol
Basham, Mark
French, Andrew P.
author_facet Luengo, Imanol
Basham, Mark
French, Andrew P.
author_sort Luengo, Imanol
building Nottingham Research Data Repository
collection Online Access
description Automatic segmentation of challenging biomedical volumes with multiple objects is still an open research field. Automatic approaches usually require a large amount of training data to be able to model the complex and often noisy appearance and structure of biological organelles and their boundaries. However, due to the variety of different biological specimens and the large volume sizes of the datasets, training data is costly to produce, error prone and sparsely available. Here, we propose a novel Selective Labeling algorithm to overcome these challenges; an unsupervised sub-volume proposal method that identifies the most representative regions of a volume. This massively-reduced subset of regions are then manually labeled and combined with an active learning procedure to fully segment the volume. Results on a publicly available EM dataset demonstrate the quality of our approach by achieving equivalent segmentation accuracy with only 5 % of the training data.
first_indexed 2025-11-14T19:57:05Z
format Article
id nottingham-44843
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:57:05Z
publishDate 2016
publisher Springer Verlag
recordtype eprints
repository_type Digital Repository
spelling nottingham-448432020-05-04T18:10:16Z https://eprints.nottingham.ac.uk/44843/ Selective labeling: identifying representative sub-volumes for interactive segmentation Luengo, Imanol Basham, Mark French, Andrew P. Automatic segmentation of challenging biomedical volumes with multiple objects is still an open research field. Automatic approaches usually require a large amount of training data to be able to model the complex and often noisy appearance and structure of biological organelles and their boundaries. However, due to the variety of different biological specimens and the large volume sizes of the datasets, training data is costly to produce, error prone and sparsely available. Here, we propose a novel Selective Labeling algorithm to overcome these challenges; an unsupervised sub-volume proposal method that identifies the most representative regions of a volume. This massively-reduced subset of regions are then manually labeled and combined with an active learning procedure to fully segment the volume. Results on a publicly available EM dataset demonstrate the quality of our approach by achieving equivalent segmentation accuracy with only 5 % of the training data. Springer Verlag 2016-09-21 Article PeerReviewed Luengo, Imanol, Basham, Mark and French, Andrew P. (2016) Selective labeling: identifying representative sub-volumes for interactive segmentation. Lecture Notes in Computer Science, 9993 . pp. 17-24. ISSN 0302-9743 Unsupervised; Sub-volume proposals; Interactive segmentation; Active learning; Affinity clustering; Supervoxels https://link.springer.com/chapter/10.1007/978-3-319-47118-1_3 doi:10.1007/978-3-319-47118-1_3 doi:10.1007/978-3-319-47118-1_3
spellingShingle Unsupervised; Sub-volume proposals; Interactive segmentation; Active learning; Affinity clustering; Supervoxels
Luengo, Imanol
Basham, Mark
French, Andrew P.
Selective labeling: identifying representative sub-volumes for interactive segmentation
title Selective labeling: identifying representative sub-volumes for interactive segmentation
title_full Selective labeling: identifying representative sub-volumes for interactive segmentation
title_fullStr Selective labeling: identifying representative sub-volumes for interactive segmentation
title_full_unstemmed Selective labeling: identifying representative sub-volumes for interactive segmentation
title_short Selective labeling: identifying representative sub-volumes for interactive segmentation
title_sort selective labeling: identifying representative sub-volumes for interactive segmentation
topic Unsupervised; Sub-volume proposals; Interactive segmentation; Active learning; Affinity clustering; Supervoxels
url https://eprints.nottingham.ac.uk/44843/
https://eprints.nottingham.ac.uk/44843/
https://eprints.nottingham.ac.uk/44843/