Learning cellular morphology with neural networks

被引:42
作者
Schubert, Philipp J. [1 ]
Dorkenwald, Sven [1 ]
Januszewski, Michal [2 ]
Jain, Viren [3 ]
Kornfeld, Joergen [1 ]
机构
[1] Max Planck Inst Neurobiol, Electron Photons Neurons, D-82152 Planegg Martinsried, Germany
[2] Google AI, CH-8002 Zurich, Switzerland
[3] Google AI, Mountain View, CA 94043 USA
关键词
CHALLENGES; RESOLUTION;
D O I
10.1038/s41467-019-10836-3
中图分类号
O [数理科学和化学]; P [天文学、地球科学]; Q [生物科学]; N [自然科学总论];
学科分类号
07 ; 0710 ; 09 ;
摘要
Reconstruction and annotation of volume electron microscopy data sets of brain tissue is challenging but can reveal invaluable information about neuronal circuits. Significant progress has recently been made in automated neuron reconstruction as well as automated detection of synapses. However, methods for automating the morphological analysis of nanometer-resolution reconstructions are less established, despite the diversity of possible applications. Here, we introduce cellular morphology neural networks (CMNs), based on multi-view projections sampled from automatically reconstructed cellular fragments of arbitrary size and shape. Using unsupervised training, we infer morphology embeddings (Neuron2vec) of neuron reconstructions and train CMNs to identify glia cells in a supervised classification paradigm, which are then used to resolve neuron reconstruction errors. Finally, we demonstrate that CMNs can be used to identify subcellular compartments and the cell types of neuron reconstructions.
引用
收藏
页数:12
相关论文
共 46 条
  • [1] [Anonymous], P IEEE INT C COMP VI
  • [2] [Anonymous], 2018, IEEE Trans. Pattern Anal. Mach. Intell.
  • [3] [Anonymous], MORPHOLOGICAL ERROR
  • [4] [Anonymous], 2014, AUTOMATIC NEURON TYP
  • [5] [Anonymous], 2014, VERY DEEP CONVOLUTIO
  • [6] [Anonymous], 2015, P IEEE C COMPUTER VI, DOI 10.1109/CVPR.2015.7298801
  • [7] [Anonymous], 2015, PROC CVPR IEEE
  • [8] [Anonymous], 2015, INT C LEARNING REPRE
  • [9] [Anonymous], 2017, NEURAL INFORMATION P
  • [10] [Anonymous], INT C MED IM COMP CO