AxonEM Dataset: 3D Axon Instance Segmentation of Brain Cortical Regions

Donglai Wei1,†  Kisuk Lee2,†  Hanyu Li3  Ran Lu2  J. Alexander Bae2  Zequan Liu4,*  Lifu Zhang5,*  Márcia dos Santos6,*  Zudi Lin1  Thomas Uram7  Xueying Wang1     Ignacio Arganda-Carreras8,9,10     Brian Matejek1  Narayanan Kasthuri3,7     Jeff W. Lichtman1     Hanspeter Pfister1    
1Harvard University     2Princeton University     3University of Chicago     4RWTH Aachen University     5Boston University
6Universidade do Vale do Rio dos Sinos     7Argonne National Laboratory     8Donostia International Physics Center (DPIC)
9University of the Basque Country (UPV/EHU)     10Ikerbasque, Basque Foundation for Science
† Equal contribution.
* Works were done as interns at Harvard University.
MICCAI 2021 / Grand Challenge

[Paper ]      [Code]      [Dataset]


     Electron microscopy (EM) enables the reconstruction of neural circuits at the level of individual synapses, which has been transformative for scientific discoveries. However, due to the complex morphology, an accurate reconstruction of cortical axons has become a major challenge. Worse still, there is no publicly available large-scale EM dataset from the cortex that provides dense ground truth segmentation for axons, making it difficult to develop and evaluate large-scale axon reconstruction methods. To address this, we introduce the AxonEM dataset, which consists of two 30x30x30 μm3 EM image volumes from the human and mouse cortex, respectively. We thoroughly proofread over 18,000 axon instances to provide dense 3D axon instance segmentation, enabling large-scale evaluation of axon reconstruction methods. In addition, we densely annotate nine ground truth subvolumes for training, per each data volume. With this, we reproduce two published state-of-the-art methods and provide their evaluation results as a baseline. We publicly release our code and data here to foster the development of advanced methods.


(Left) We plot the distribution of the axon instance length for AxonEM and SNEMI3D datasets. (Right) We show the 3D rendering of neurons with somas in AxonEM dataset. Two large glial cells (not rendered) in AxonEM-H occupy the space between the blue and pink neurons, leading to much fewer long axons compared to AxonEM-M.There is a strong linear correlation between the volume and length mitochondria in both volumes, which is the average thickness of the instance. While the MitoEM-H has more small instances, the MitoEM-R has more large instances with complex morphologies.


      title={AxonEM Dataset: 3D Axon Instance Segmentation of Brain Cortical Regions}, 
      author={Donglai Wei and Kisuk Lee and Hanyu Li and Ran Lu and J. Alexander Bae and Zequan Liu and 
      Lifu Zhang and Márcia dos Santos and Zudi Lin and Thomas Uram and Xueying Wang and Ignacio 
      Arganda-Carreras and Brian Matejek and Narayanan Kasthuri and Jeff Lichtman and Hanspeter Pfister},


KL, RL, and JAB were supported by the Intelligence Advanced Research Projects Activity (IARPA) via Department of Interior/ Interior Business Center (DoI/IBC) contract number D16PC0005, NIH/NIMH (U01MH114824, U01MH117072, RF1 MH117815), NIH/NINDS (U19NS104648, R01NS104926), NIH/NEI (R01EY027 036), and ARO (W911NF-12-1-0594), and are also grateful for assistance from Google, Amazon, and Intel. DW, ZL, JL, and HP were partially supported by NSF award IIS-1835231. I. A-C would like to acknowledge the support of the Beca Leonardo a Investigadores y Creadores Culturales 2020 de la Fundación BBVA. We thank Viren Jain, Michał Januszewski and their team for generating the initial segmentation for AxonEM-H, and Daniel Franco-Barranco for setting up the challenge using AxonEM.

Declaration of Interests

KL, RL, and JAB disclose financial interests in Zetta AI LLC.