Self-Supervised Learning for Biological Sample Localization in 3D Tomographic Images

TL;DR

This paper introduces self-supervised learning methods for efficient biological sample localization in 3D tomographic images, significantly reducing storage and improving localization accuracy in high-throughput CT setups.

Contribution

It proposes two novel self-supervised loss functions and utilizes uncertainty estimation for accurate sample localization in biological CT images.

Findings

Achieves less than 1.5% relative error in sample localization.

Reduces storage requirements by a factor of four.

One loss functions effectively as a pre-training task for segmentation.

Abstract

In synchrotron-based Computed Tomography (CT) there is a trade-off between spatial resolution, field of view and speed of positioning and alignment of samples. The problem is even more prominent for high-throughput tomography--an automated setup, capable of scanning large batches of samples without human interaction. As a result, in many applications, only 20-30% of the reconstructed volume contains the actual sample. Such data redundancy clutters the storage and increases processing time. Hence, an automated sample localization becomes an important practical problem. In this work, we describe two self-supervised losses designed for biological CT. We further demonstrate how to employ the uncertainty estimation for sample localization. This approach shows the ability to localize a sample with less than 1.5\% relative error and reduce the used storage by a factor of four. We also show that one of the proposed losses works reasonably well as a pre-training task for the semantic segmentation.