Robust Registration of Calcium Images by Learned Contrast Synthesis

TL;DR

This paper introduces a machine learning approach to contrast synthesis that significantly improves multi-modal calcium image registration, reducing failure rates in in-vivo Drosophila brain imaging.

Contribution

The study demonstrates that learned contrast synthesis enhances registration robustness, especially in challenging multi-modal scenarios where traditional cost functions fail.

Findings

Registration failure rate reduced from 40% to 15%.

Contrast synthesis improves registration accuracy.

Method effective for in-vivo Drosophila brain images.

Abstract

Multi-modal image registration is a challenging task that is vital to fuse complementary signals for subsequent analyses. Despite much research into cost functions addressing this challenge, there exist cases in which these are ineffective. In this work, we show that (1) this is true for the registration of in-vivo Drosophila brain volumes visualizing genetically encoded calcium indicators to an nc82 atlas and (2) that machine learning based contrast synthesis can yield improvements. More specifically, the number of subjects for which the registration outright failed was greatly reduced (from 40% to 15%) by using a synthesized image.