![]() |
Reading coordinates from SMI iView ETG application
Dear Reader,
We have this iView eyetracking device. It is a USB device, it is not wireless. There is a server program that is capable of correctly displaying the output of the glasses. This program is called iView ETG. Vizard, namely python, is probably way too slow to do any real-time video channeling or processing. This is not the issue, we are most likely satisfied if we would just be able to obtain the gaze direction. Originally, I wrote a small program based on the vizard help file. I later replaced this program with a program found on this forum. Neither of the programs worked. I think it might have something to do with the fact that I was using the iView X interface instead of the iView NG interface, but actually, I have no clue. When I connect to 127.0.0.1 at port 4444, the program does not produce an error message. According to the manual of the iView NG SDK, the default port would be 1100 which is equivalent to using 0, although this option clearly does not work. I have written an event that prints the gaze direction. The event gets called each frame by using a callback. I use the function tracker.getPosition in order to print the gaze direction. You can choose one of two constants: smi.PIXELS and smi.NORMALIZED. Both constants result in invalid output. The latter may have something to do with the fact that the function tracker.getCalibrationSize() returns the pair (0,0). In this way, it does seem to make some sense that the output would be scaled to absolute 0. However, I cannot see how to enlarge the calibration size such that the output would lie within a 2D domain. Can someone help me with this issue? Thank you. Best regards, Wouter Buddingh (using the forum account of Sjaak Verwaaijen) |
I'll need to check with a developer. Have you also tried connecting to data streaming from the iView X™ software?
|
To Jeff,
SMI said that I would not require the iView X application. They therefore did not supply me with a licence key for the program. It could be the case that it might just not be possible. The manual also talks about iView X, and not iView ETG. This is really odd though, because Worldviz and SMI are partners, you would expect the stuff to run flawlessly. Any other thoughts? Thank you. p.s. We recently bought the glasses. They were very expensive (about 10.000 euro's) you would expect that such a system would operate without any trouble. The iView ETG application clearly works, but I would also expect that the interface to vizard would work without any trouble. The manual supplied within the iView SDK isn't totally finished. It only contains C++ interface examples. The python interface examples are left blank. By the way, there is a difference between python and Vizard python, although I think that the python code should also work within vizard, probably not the other way around. |
Any updates here? I have the same question.
|
You can use VRPN to stream data from iView ETG to Vizard. The SMI downloads page has a link for a VRPN server. In Vizard, use code like the following to get a handle to the trackers:
Code:
vrpn = viz.add('vrpn7.dle') |
Thanks, Jeff
|
All times are GMT -7. The time now is 07:35 AM. |
Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright 2002-2023 WorldViz LLC