Add IntersonSDKCxx (Interson SDK 1.X) support. about pluslib HOT 1 OPEN

2014-05-26 23:38
Matt McCormick
re #892: Add IntersonSDKCxx BMode support.
Committed to: plus
The vtkIntersonSDKCxxVideoSource is added with B-Mode output along with the
Test.

PlusLib support for IntersonSDKCxx is enabled with the

PLUS_USE_INTERSONSDKCXX_VIDEO

CMake configuration variable.
In plus:3330

2014-05-27 01:07
Matt McCormick
re #892: Add IntersonSDKCxx to PlusDeviceFactory.
Committed to: plus
In plus:3332

2014-05-30 13:44
Matt McCormick
re #892: Export IntersonSDKCxx_DIR in PlusLibConfig.cmake.
Committed to: plus
In plus:3348

2014-06-08 21:46
Matt McCormick
re #892: Add missing IntersonSDKCxxVideoSourceTest config file.
Committed to: plus
In plus:3389

2014-06-08 22:21
Matt McCormick
re #892: Add vtkIntersonSDKCxxVideoSource sources.
Committed to: plus
In plus:3391

2014-06-09 09:11
Andras Lasso
@matt.mccormick Thanks for your contribution! As there are several similarities between vtkIntersonSDKCxxVideoSource and vtkIntersonVideoSource it would make sense to create an abstract base class and put SDK-specific details (iSDK and SDKCxx) into two child classes. What do you think?

What are your plans with vtkIntersonSDKCxxVideoSource? Do you plan to add more features, clean up things, etc? Do you have time to contribute with implementation and/or testing to the consolidation of the iSDK and SDKCxx classes?

2014-06-09 12:04
Matt McCormick
@lassoan I did not expect much overlap with the old SDK, but I am open to suggestions. Interson said they re-wrote the new SDK from scratch, so I do not expect many similarities. Are there particular methods you think should be shared? I don't have access to the old SDK or an old probe, so I would need some help testing.

Other todo's here include fixing the probe initialization -- there is something odd going on in that the SDK example application needs to be run before it starts populating the buffer. RF output needs to be added, too. Any suggestions on examples to look at for reference on how to best add both BMode and RF dual output are welcome :-).

2014-06-09 12:52
Andras Lasso
These Interson classes are quite similar. The hardware architecture and capabilities are also very similar, so it is very likely that we'll use them the same way.

I'll create the common base class and move both classes into the same directory. I just wait for confirmation from Guillermo that he doesn't have any pending changes.

We'll have a look at concurrent B/RF-mode acquisition. I think the BK ultrasound class may serve as example.

2014-06-09 12:54
Andras Lasso
A short note on this comment (in SetDepth): // Since we are not using their scan converter, this does not have much effect

It's not just the scan conversion parameters, but pulse frequency and sample rate parameters need to be changed, too (see the attached ppt in #869), if you want to have optimal image quality. Also, if we want to do the scan conversion then we need to get the probe geometry from the SDK (or define it in the config file).

2014-06-09 13:15
Matt McCormick
The pulse frequency / sample rate / clock divider no longer need to be paired like they did in the old probe / SDK.

They don't export the probe geometry in the public methods of their SDK, so we will either need to hard code them on in the Plus .cxx files or define them in the config file.

2014-06-09 17:10
Matt McCormick
re #892: Updates to vtkIntersonSDKCxxVideoSource.
Committed to: plus
In plus:3393

2014-06-09 17:11
Andras Lasso
Is the frequency/sample rate/clock divider constant for all depths when the new SDK is used?

If the probe geometry is not available in the SDK then we should store it in the config file (there is already an XML element to describe probe geometry).

2014-06-09 17:12
Andras Lasso
Do you plan to add IntersonSDKCxx to PlusBuild? (add an option, if enabled then download from github, build it, make PlusLib to use it - the same way as for OpenIGTLink, ITK, VTK)

2014-06-09 17:17
Matt McCormick
With the new probe/SDK, yes, the sample rate/clock divider is constant for all frequencies. When a different depth is set on their scan converter, it just shows the depths requested (a subset of what is actually acquired).

I'll add the probe geometry to the config file, then.

Yes, I'll add IntersonSDKCxx to PlusBuild like the other projects. I'll get a config file that does the scan conversion so we can view the probe content with PlusApp.

Thanks for the pointers.

2014-06-11 20:43
Matt McCormick
re #892: Initialize Interson SDK scan converter.
Committed to: plus
Required to initialize the probe even if we do not use their scan converter.
In plus:3410

2014-06-18 22:48
Matt McCormick
re #892: Add IntersonSDKCxx ExternalProject.
Committed to: plus
PLUS_USE_INTERSONSDKCXX_VIDEO in PlusBuild will build the required library.
In plus:3446

2014-06-18 23:56
Matt McCormick
re #892: Add PlusDeviceSet_Server_IntersonSDKCxx_GP3.5.xml start.
Committed to: plus
Still needs scan conversion.
In plus:3447

2014-06-20 02:18
Matt McCormick
re #892: IntersonSDK BMode scan conversion.
Committed to: plus
In plus:3450

2014-06-22 17:30
Matt McCormick
re #892: IntersonSDKCxx: do not downcast to Curvilinear.
Committed to: plus
The vtkUsScanConvert interface can be used for all transducer.

Suggested-by: Andras Lasso.
In plus:3453

2014-06-22 17:49
Matt McCormick
re #892: IntersonSDKCxx: Do not downcast scan converter.
Committed to: plus
In plus:3454

2014-06-22 20:47
Matt McCormick
re #892: IntersonSDKCxx: start RF collection.
Committed to: plus
Test currently crashes.
In plus:3456

2014-06-23 16:34
Matt McCormick
re #892: vtkPlusDevice::RequestData use correct ScalarType.
Committed to: plus
Use the ScalarType from the vtkPlusBuffer in the vtkPlusDevice::RequestData
before AllocateOutputs. Previously, it was hard coded to unsigned char, which
changed the ScalarType for the output with signed short RF data.
In plus:3458

2014-06-26 16:59
Matt McCormick
re #892: Improve plotting parameters for IntersonSDK RF test.
Committed to: plus
In plus:3469

2014-06-27 18:35
Matt McCormick
re #892: IntersonSDK RF -> BMode generation.
Committed to: plus
Support two output channels, one with the RF data and the other with the
BMode image simultaneously generated from the RF.
In plus:3470

2014-06-27 18:37
Matt McCormick
re #892: Run IntersonSDK RF to BMode test.
Committed to: plus
In plus:3471

2014-06-27 18:52
Matt McCormick
re #892: Add Interson configuration for collecting RF.
Committed to: plus
Using fCal, it displays the B-Mode generated from the RF and writes the RF to a
MetaImage file.
In plus:3472

2014-06-27 19:48
Andras Lasso
Very nice!

2014-06-28 15:07
Matt McCormick
Plus BMode creation looks much nicer than the on-board BMode :-)

2014-06-28 16:12
Matt McCormick
re #892: Set Interson acquisition frame rate.
Committed to: plus
In plus:3473

2014-06-28 16:26
Matt McCormick
re #892: Remove DepthMm option from IntersonSDK.
Committed to: plus
This is for the SDK scan converter, which we do not use.
In plus:3474

2014-06-28 17:18
Matt McCormick
re #892: Make IntersonSDK PulseVoltage configurable.
Committed to: plus
In plus:3475

2014-06-29 17:01
Matt McCormick
re #892: Add RfDecimation option.
Committed to: plus
Cuts sampling rate in half, so depth doubles.

Also fix configuration file depths.
In plus:3476

2014-06-29 18:03
Matt McCormick
re #892: Reduce IntersonSDK time to start.
Committed to: plus
In plus:3477

2014-06-29 23:17
Andras Lasso
@matt.mccormick Plus should provide hardware abstraction for US imaging devices as much as possible. For most US systems the desired imaging depth can be set and Plus and/or the device driver figures out the optimal acquisition parameters to achieve that. Could you implement such mechanism for this device, too?

For example:
Instead of requiring the user to figure out if he needs decimation or not, we could decide automatically from the requested depth (we would only enable decimation if the desired depth cannot be reached without decimation) and we might also clip the part of the RF lines that are over the requested depth (to optimize speed and memory consumption).
It would be nice if the scan conversion parameters could be computed from a few simple user inputs, such as: DepthMm and TransducerModel (a string, such as "GP 3.5MHz", which would be used to look up/compute the optimal TransducerGeometry, RadiusStartMm, ThetaStartDeg, ThetaStopDeg, OutputImageSizePixel, TransducerCenterPixel, and OutputImageSpacingMmPerPixel values). The SDK should already provide at least the TransducerModel name (so the user would not need to specify it), but ideally the SDK should also provide info about the probe's physical dimensions and capabilities. If the SDK does not provide these info then we should ask Interson to add such methods (and until they add it we may need to ask the user to specify the TransducerModel with a string and add tables to Plus that contain parameters for each TransducerModel).

2014-06-30 00:10
Matt McCormick
@lassoan Yes, it sounds good to provide or compute reasonable defaults when possible. I'll add this to the roadmap, although I will need to work on other priorities in the near term.

2014-11-23 22:17
Matt McCormick
re #892: vtkIntersonSDKCxxVideoSourceTest compatible with VTK 6.
Committed to: plus
In plus:3858

2014-11-23 22:32
Matt McCormick
re #892: Doc IntersonSDK needs Interson.dll instead of WDAPI1010.dll.
Committed to: plus
In plus:3859

2015-06-03 22:17
Matt McCormick
re #892: Fix IntersonSDKCxx image orientations.
Committed to: plus
The Plus documentation now clearly states the output image orientation of all
B-mode images must be MF and RF must be FM.
In plus:4176

2015-08-07 09:44
Matt McCormick
re #892: Update the IntersonSDKCxx configuration orientations.
Committed to: plus
In plus:4239

2015-08-07 09:52
Matt McCormick
re #892: Revert Interson orientation changes.
Committed to: plus
They were not correct.
In plus:4240

2015-08-07 10:41
Matt McCormick
re #892: Increase startup delay for Interson Rf Acquision.
Committed to: plus
Avoid error of VTK render window not available.
In plus:4241

from pluslib.