Semi-Supervised Segmentation of Mitochondria from Electron Microscopy Images Using Spatial Continuity

TL;DR

This paper introduces a semi-supervised deep learning approach for mitochondrial segmentation in electron microscopy images, leveraging spatial continuity and data augmentation to reduce annotation requirements while maintaining high accuracy.

Contribution

The study presents a novel semi-supervised model that achieves comparable performance to fully supervised methods using only about 20% of the annotated data, utilizing spatial continuity and morphological augmentation.

Findings

Achieves similar accuracy to fully supervised models with 80% less annotated data.

Utilizes spatial continuity and morphological augmentation for improved segmentation.

Demonstrates versatility in segmenting other spatially continuous structures.

Abstract

Morphology of mitochondria plays critical roles in mediating their physiological functions. Accurate segmentation of mitochondria from 3D electron microscopy (EM) images is essential to quantitative characterization of their morphology at the nanometer scale. Fully supervised deep learning models developed for this task achieve excellent performance but require substantial amounts of annotated data for training. However, manual annotation of EM images is laborious and time-consuming because of their large volumes, limited contrast, and low signal-to-noise ratios (SNRs). To overcome this challenge, we propose a semi-supervised deep learning model that segments mitochondria by leveraging the spatial continuity of their structural, morphological, and contextual information in both labeled and unlabeled images. We use random piecewise affine transformation to synthesize comprehensive and realistic mitochondrial morphology for augmentation of training data. Experiments on the EPFL dataset show that our model achieves performance similar as that of state-of-the-art fully supervised models but requires only ~20% of their annotated training data. Our semi-supervised model is versatile and can also accurately segment other spatially continuous structures from EM images. Data and code of this study are openly accessible at https://github.com/cbmi-group/MPP.