Presentation Information for DICOM Printing

[vijaya]

As per my current understanding of specifying presentation information to SCP, the SCU can either transmit pre-formatted image in p-value or send pre-formatted bitmap with a LUT so that print server can convert bitmap to p-values.

I have the following queries:

1. Are the LUTs standardised and if so where can we get information about which LUT to use. Does the type of LUT to be used change with the
image type?
2. If image grayscale or color values (RGB) values are mapped to LUT values, how is the p-value derived from this mapping?
3. Finally, how does the SCU convert p-values to Optical density values for printing?
4. If Presentation LUT is not supported, as it is an optional SOP class, how will the p-values be sent by the SCU?

[Marco Eichelberg]

Why are you cross-posting this question both to this forum and to comp.protocols.dicom (and not even mentioning that)? I perceive that as somewhat disrespectful to the readers of one forum who will not see the discussion in the other one. This answer has also been posted to comp.protocols.dicom where this discussion really belongs since it is not related in any way to DCMTK or DCMPRINT.

1. Are the LUTs standardised and if so where can we get information about which LUT to use. Does the type of LUT to be used change with the image type?

The syntax of the LUT is standardised (i.e. possible numbers of entries, bits per entry etc. are defined in the N-CREATE-RQ primitive for the Presentation LUT SOP Class). The LUT content is not standardized - an SCU may send whatever may seem appropriate.

2. If image grayscale or color values (RGB) values are mapped to LUT values, how is the p-value derived from this mapping?

Presentation LUTs can only be used with monochrome images (Basic Grayscale Print Management). The result of mapping the Basic Grayscale Image Box pre-formatted pixel values through the Presentation LUT produces p-values by definition - i.e. whatever the result of the LUT is, the printer will interpret that as p-values.

3. Finally, how does the SCU convert p-values to Optical density values for printing?

That is defined in great detail in part 14 of the DICOM standard. The transformation is based on the DICOM Grayscale Standard Display Function.

4. If Presentation LUT is not supported, as it is an optional SOP class, how will the p-values be sent by the SCU?

If presentation LUT is not supported, then the precise interpretation of the Image Box pixel values by the printer is simply undefined. The printer may apply a LUT defined in the printer configuration, or may select a LUT based on the Configuration Information attribute in the Basic Film Box (many printers do support that). The printer may interpret the raw pixel values as p-values or a linear optical density. Unless the SCU knows exactly which printer it is talking to, and how this printer is configured, the result is somewhat unpredictable, which is exactly the reason why the Presentation LUT SOP class was defined as an add-on to solve that problem.

[vijaya]

I apologise for the inconvenience caused. In future I will post my queries to comp.protocols.dicom.