Index of /DIGdata/MAX4Img_AxonMorphology/XNH datasets/axonSegmentations

[ICO]NameLast modifiedSize
[PARENTDIR]Parent Directory  -
[DIR]highResolutionFiles/2020-10-23 22:15 -
[DIR]lowResolutionFiles/2020-10-21 10:12 -
[TXT]README.html2020-10-23 22:12 2.0K

Axon Segmentations Info

The folder "AxonSegmentations" contains the following 4 entries:

  1. "highResolutionFiles/R1.vol" Dim: [2048 2048 2048], voxel size: [75 75 75] nm, Volume size (memory): ~32 GB
    X-ray Nano-Holotomography volume of the monkey brain splenium, (referred to as "region 1" or "R1" in the context of other volumes included in this data package).

  2. "highResolutionFiles/AxonVoxelIndices/*.txt"
    This directory contains 54 .txt files with the coordinates for each of the 54 axons presented in Andersson et al. 2020. See "HighResolutionFiles/AxonVoxelIndices_info.txt" for more information.

  3. "lowResolutionFiles/downsampledR1.nii" Dim: [410 410 410], voxel size: [375 375 375] nm
    Contains downsampled (factor of 5) and smoothed version of File 1. For viewing on a standard desktop computer, without running in to memory issues.

  4. "lowResolutionFiles/downsampledR1_54AxonSeg.nii" Dim: [410 410 410], voxel size: [375 375 375] nm
    Contains downsampled (factor of 5) axon segmentation. Produced solely for easy viewing in conjunction with File 3 on a standard desktop computer, with a 3D image viewer like "ITKSnap" or "ImageJ".

File 3 has been downsampled for the purpose of easy viewing and preliminary segmentation of axons (see Methods of Andersson et al 2020) to guide the high resolution segmentation of the axons (provided here as voxel indices). To produce File 3, File 1 was downsampled to a size of 410 x 410 x 410 by extracting every 5th slice, then implementing a slice-wise downsampling by cubic interpolation and applying a Gaussian smoothing kernel (size 5).

File 4 is also downsampled by a factor of 5 to enable easy viewing of axonal segmentations on a desktop computer. This is done by downsampling the full [2048 2048 2048] axon segmentation volume to a size of [410 410 410] voxels using "nearest neighbour" interpolation.