Large files (~20MB) are auto-scaling to 0-99 grayscale?

Support
, ,
1

Hi! Original DICOM files are being loaded into slicer to be analyzed for hard tissue density analysis (bone, dentin, enamel, cementum). Because the mineral densities are so dense and difficult to distinguish at lower resolutions (10µm), we have to scan them at 7µm, which increases the size of the files. The 10µm scans can’t distinguish these mineral types, but the 7µm scans can. Oddly, the 10µm scans have a large grayscale range during segmentation (-1000 to +1500), but the 7µm scans seem to autoscale (0 to 99). Would anyone be able to please help with this? Thanks!

2

Slicer loads images the same way regardless of the voxel size. If you find a difference between how these images appear in Slicer then it is due to that the operator used different settings on the scanner, or the images were reconstructed or exported differently.

3

Thanks! The files were treated the same though – same settings and reconstruction…

4

Ranges are so different it is unlikely the same reconstruction protocol. Slicer uses the reported intensities in the volume. At the minimum exported datasets are different data type (signed 16 bit for your first data, and unsigned 8 bit for the higher resolution one).

You really should reach out to the lab or the person that did the reconstruction.

5

Thanks! Just reached out to the lab and this is the response I got, so I think it may be something that’s happening when loading into slicer?:

"The raw data is reconstructed automatically by the Scanco programme. They share the same settings.
The original data point values in Scanco are 8 bit signed integers (Please see attached note). However, this could be changed when you import to other software programmes and export. "

Values_Threshold_Density_Hounfield Units.png
Values_Threshold_Density_Hounfield Units.png951×938 78.7 KB

6

nope, slicer does not modify intensity values. However, I don’t think Scanco outputs DICOMs, it has its own format called AIM (which you can read directly into Slicer actually). Whoever is doing the DICOM conversion of these files, is changing the intensities.

7

Hmm weird…both 7um and 10um dicom file exports from scanco upload to slicer, but the 10um intensity doesn’t change, just the 7um does…? it’s a mystery!

8

I would ask for the aim files as generated by the Scanco, not the dicoms. AIM format is readable by slicer.