102 — ShuTu: Open-Source Software for Efficient and Accurate Reconstruction of Dendritic Morphology

ShuTu is a software package that automatically generates reconstructions of neurons in 3D space from bright-field microscopy. The system then allows for manual proofreading and editing in order to improve the quality of the annotations.

One benefit of bright-field microscopy is that there is no need to align the images after scanning — unlike slicing systems like electron microscopy, the ZEN blue microscope outputs one large 3D volume file. However, it is still necessary to stitch along the X and Y axes.

The authors present their image segmentaton algortihms and generate a SWC file from a binary mask, which can then be edited manually by a human.

In addition to the technical challenges themselves, this paper is interesting to me because of the pipelining of workflows to modify and edit machine-generated annotations. It’s super interesting to see how teams manage human annotation: While computation can be expensive, human labor, at least per-hour, is way more expensive. Optimizing these two costs is a difficult problem, and protocols like this one uses ShuTu to make the problem simpler.