SSHMT: Semi-supervised Hierarchical Merge Tree for Electron Microscopy Image Segmentation
Published in
European Conference on Computer Vision(2016)External Links:
TL
TL;DR
A Bayesian model is proposed that combines the supervised and the unsupervised information for probabilistic learning and consistently performs close to the state-of-the-art supervised method with the full labeled data set and significantly outperforms the supervised method with the same labeled subset.
Abstract
Region-based methods have proven necessary for improving segmentation accuracy of neuronal structures in electron microscopy (EM) images. Most region-based segmentation methods use a scoring function to determine region merging. Such functions are usually learned with supervised algorithms that demand considerable ground truth data, which are costly to collect. We propose a semi-supervised approach that reduces this demand. Based on a merge tree structure, we develop a differentiable unsupervised loss term that enforces consistent predictions from the learned function. We then propose a Bayesian model that combines the supervised and the unsupervised information for probabilistic learning. The experimental results on three EM data sets demonstrate that by using a subset of only 3% to 7% of the entire ground truth data, our approach consistently performs close to the state-of-the-art supervised method with the full labeled data set, and significantly outperforms the supervised method with the same labeled subset.
References29 items
1
The Morphological Approach to Segmentation: The Watershed Transformation
2
Efficient Classifier Training to Minimize False Merges in Electron Microscopy Segmentation
3
Crowdsourcing the creation of image segmentation algorithms for connectomics
4
Learning to Segment: Training Hierarchical Segmentation under a Topological Loss
5