MiShape: 3D Shape Modelling of Mitochondria in Microscopy

TL;DR

MiShape is a generative model that learns mitochondrial shapes from electron microscopy data, enabling accurate 3D shape reconstruction from limited fluorescence microscopy images and facilitating realistic dataset generation for segmentation tasks.

Contribution

This work introduces MiShape, a novel implicit shape prior model for mitochondria that improves 3D reconstruction and dataset simulation from fluorescence microscopy images.

Findings

MiShape accurately reconstructs 3D mitochondrial shapes from 2D images.

The model generates realistic mitochondrial shapes for dataset augmentation.

MiShape enhances segmentation and transformation tasks in microscopy imaging.

Abstract

Fluorescence microscopy is a quintessential tool for observing cells and understanding the underlying mechanisms of life-sustaining processes of all living organisms. The problem of extracting 3D shape of mitochondria from fluorescence microscopy images remains unsolved due to the complex and varied shapes expressed by mitochondria and the poor resolving capacity of these microscopes. We propose an approach to bridge this gap by learning a shape prior for mitochondria termed as MiShape, by leveraging high-resolution electron microscopy data. MiShape is a generative model learned using implicit representations of mitochondrial shapes. It provides a shape distribution that can be used to generate infinite realistic mitochondrial shapes. We demonstrate the representation power of MiShape and its utility for 3D shape reconstruction given a single 2D fluorescence image or a small 3D stack of 2D slices. We also showcase applications of our method by deriving simulated fluorescence microscope datasets that have realistic 3D ground truths for the problem of 2D segmentation and microscope-to-microscope transformation.