A deep learning-based strategy for producing dense 3D segmentations from sparsely annotated 2D images

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: June 15, 2024

ABSTRACT Producing dense 3D reconstructions from biological imaging data is a challenging instance segmentation task that requires significant ground-truth training for effective and accurate deep learning-based models. Generating intense human effort to annotate each of an object across serial section images. Our focus on the especially complicated brain neuropil, comprising extensive interdigitation dendritic, axonal, glial processes visualized through electron microscopy. We developed novel method generate segmentations rapidly sparse 2D annotations few objects single sections. Models trained generated achieved similar accuracy as those expert annotations. Human time was reduced by three orders magnitude could be produced non-expert annotators. This capability will democratize generation large image volumes needed achieve circuits measures circuit strengths.

Language: Английский

Beyond neurons: computer vision methods for analysis of morphologically complex astrocytes

Frontiers in Computer Science, Journal Year: 2024, Volume and Issue: 6

Published: Sept. 25, 2024

The study of the geometric organization biological tissues has a rich history in literature. However, geometry and architecture individual cells within traditionally relied upon manual or indirect measures shape. Such rudimentary are largely result challenges associated with acquiring high resolution images cellular components, as well lack computational approaches to analyze large volumes high-resolution data. This is especially true brain tissue, which composed complex array cells. Here we review tools that have been applied unravel nanoarchitecture astrocytes, type cell increasingly being shown be essential for function. Astrocytes among most structurally functionally diverse mammalian body partner neurons. Light microscopy does not allow adequate astrocyte morphology, however, large-scale serial electron data, provides nanometer 3D models, enabling visualization fine, convoluted structure astrocytes. Application computer vision methods resulting nanoscale models helping reveal organizing principles but complete understanding its functional implications will require further adaptation existing tools, development new approaches.

Language: Английский

Global Neuron Shape Reasoning with Point Affinity Transformers

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 24, 2024

Connectomics is a subfield of neuroscience that aims to map the brain's intricate wiring diagram. Accurate neuron segmentation from microscopy volumes essential for automating connectome reconstruction. However, current state-of-the-art algorithms use image-based convolutional neural networks are limited local shape context. Thus, we introduce new framework reasons over global with novel point affinity transformer. Our embeds (multi-)neuron cloud into fixed-length feature set which can decode any pair affinities, enabling clustering clouds automatic proofreading. We also show learned easily be mapped contrastive embedding space enables type classification using simple KNN classifier. approach excels in two demanding connectomics tasks: proofreading errors and classifying types. Evaluated on three benchmark datasets derived connectomes, our method outperforms transformers, graph networks, unsupervised baselines.

Language: Английский