Hi @neheller, image intensity range is negative to positive (for example -1024 to 2500). I think strange normalization is done at the time to save nifty (please correct and clarify here).
In addition to the above, slice thickness is also different for each image. right?
The intensity values are raw Hounsfield Units which are a measure of X-ray attenuation. The standard practice is to normalize them so that water is 0 HU and air at STP is -1000 HU. This normalization is done at the time that the scan is captured, and we did not adjust these HU values in any way during conversion.
For the slice thickness, yes there’s a fairly wide range of slice thicknesses. These scans were collected retrospectively from routine care, so it should be a fairly good approximation of the true distribution of the slice thicknesses chosen for this indication.