In this work we present a novel approach for unsupervised multi-graph matching, which applies to problems for which a Gaussian distribution of keypoint features can be assumed. We leverage cycle consistency as loss for self-supervised learning, and determine Gaussian parameters through Bayesian Optimization, yielding a highly efficient approach that scales to large datasets. Our fully unsupervised approach enables us to reach the accuracy of state-of-the-art supervised methodology for the biomedical use case of semantic cell annotation in 3D microscopy images of the worm C. elegans. To this end, our approach yields the first unsupervised atlas of C. elegans, i.e. a model of the joint distribution of all of its cell nuclei, without the need for any ground truth cell annotation. This advancement enables highly efficient semantic annotation of cells in large microscopy datasets, overcoming a current key bottleneck. Beyond C. elegans, our approach offers fully unsupervised construction of cell-level atlases for any model organism with a stereotyped body plan down to the level of unique semantic cell labels, and thus bears the potential to catalyze respective biomedical studies in a range of further species.

Multi-graph matching is an important problem in computer vision. Our task comes from bioimaging, where a set of 100 3D-microscopic images of worms have to be brought into correspondence. Surprisingly, virtually all existing methods are not applicable to this large-scale, real-world problem since they either assume a complete or dense problem setting, and they have so far only been applied to small-scale, toy or synthetic problems. Despite claims in literature that methods addressing complete multi-graph matching are applicable in an incomplete setting, our first contribution is to prove that their runtime would be excessive and impractical. Our second contribution is a new method for incomplete multi-graph matching that applies to real-world, larger-scale problems. We experimentally show that for our bioimaging application we are able to attain results in less than two minutes, whereas the only competing approach requires at least half an hour while producing far worse results. Furthermore, even for small-scale, dense or complete problem instances we achieve results that are at least on par with the leading methods, but an order of magnitude faster.

SUMMARY: High-resolution imaging of model organisms allows the quantification of important physiological measurements. In the case of fish with transparent embryos, these videos can visualize key physiological processes, such as heartbeat. High throughput systems can provide enough measurements for the robust investigation of developmental processes as well as the impact of system perturbations on physiological state. However, few analytical schemes have been designed to handle thousands of high-resolution videos without the need for some level of human intervention. We developed a software package, named FEHAT, to provide a fully automated solution for the analytics of large numbers of heart rate imaging datasets obtained from developing Medaka fish embryos in 96-well plate format imaged on an Acquifer machine. FEHAT uses image segmentation to define regions of the embryo showing changes in pixel intensity over time, followed by the classification of the most likely position of the heart and Fourier Transformations to estimate the heart rate. Here, we describe some important features of the FEHAT software, showcasing its performance across a large set of medaka fish embryos and compare its performance to established, less automated solutions. FEHAT provides reliable heart rate estimates across a range of temperature-based perturbations and can be applied to tens of thousands of embryos without the need for any human intervention. AVAILABILITY AND IMPLEMENTATION: Data used in this manuscript will be made available on request.