Home › Forums › Software Development › Pupil Foreshortening Error
Tagged: #erp #crm #cms #ecommerce
- This topic has 5 replies, 2 voices, and was last updated 1 year ago by
Grant [Tobii].
- AuthorPosts
- 03/08/2022 at 17:19 #22308
Michael Wimmer
ParticipantHi Tobii!
I am using a HP Omnicept G2 Reverb and record pupil size data during a video game. Is the algorithm for pupil size calculation considering gaze angle-dependencies in the pupil size? If no or not completely, is there a gaze range in x and y direcion you can recommend for a save use of the metric?
05/08/2022 at 14:20 #22324Grant [Tobii]
KeymasterHi @miwi, we cover in some detail the image-based algorithm used to determine pupil size which is independent of gaze angle @
So there is no gaze range as such, but as the document indicates, it is better use % of pupil size difference rather than any absolute value.
10/08/2022 at 14:51 #22331Michael Wimmer
ParticipantHi, thank you for your answer. Let me specify my question a little bit:
The data I collected suggest that at the moment when a user changes his gaze direction, that also the pupil size changes. Could it be that the measured pupil size changes just because of people looking around?
16/08/2022 at 20:33 #22339Grant [Tobii]
KeymasterHi @miwi, yes pupil size can vary significantly between dark and light but also there is constant subtle variation when scanning an environment. However, if the pupil size is changing much more than expected, it may be worth looking at environmental conditions during recording such as ambient lighting, spotlights, etc.
17/08/2022 at 12:09 #22340Michael Wimmer
ParticipantHi,
Thank you! However, I could see the dependency of the pupil size on the view angle (especially in X direction) even in completely controlled, constant conditions (only looking to the left/right with a black background). Recorded variations of the pupil size due to these saccades are around +/-0.2mm. Are you aware of such behavior?
12/09/2022 at 11:27 #22403Grant [Tobii]
KeymasterHi, After consulting internally, I’m afraid we do not have any particular metrics to demonstrate the relationship between gaze angle and reported pupil size. It may be worth posting on HP own forum in case another user there has performed a similar analysis.
- AuthorPosts
- You must be logged in to reply to this topic.