Hello, I am looking into some litterature about the 3D eye model, the one implemented in the Pupil Software does not exactly suit me but I would be interested if the method has been published like the pupil detection one

Ok I found the Dierkes, Kassner & Bulling paper from 2018, I guess this is the latest one?

@user-92dca7 please could you provide some links to @user-bd800a ?

Hi, we recently published the following paper which made use of the pupil labs eye tracker - I wanted to add it to your pupil citation list, not sure of how I could do this: https://www.sciencedirect.com/science/article/abs/pii/S1053811919309140

Hi @user-e63a0a thanks for the update. I will add this to the citation list.

Thanks @wrp! 🙂

Hello all,

Disregard above... Hello all, I am planning on using the Pupil labs software for a research project, and I have been following the DIY intructions from the Pupil labs website. I have a few questions for the commuinty. So far, I have trouble with the webcams. I keep getting a time stamp error...Does anyone know how I can fix this? I don't have the error infront of me but I can get the exact message another time. I am using Ubuntu 18.04 kernel version 5.0.0. Also, I have had trouble finding the correct uvc compliant webcams. For those that have made their own devices, what webcam models have you had success with with? Much appreciated.

@user-860618 I think you can ignore that error. Do you get images?

@mpk I able to to the calibration routine and record video then view it in the player. I am being guided by my professor whose said look into; He has extensive experience with computers (programming, writing programs etc) so I just assumed it was an issue by authority.

do*

@user-860618 I've been using the Logitech C525 webcam for a little while now and it's been working great. I use the Pupil Labs eye cameras, with the C525 as the world camera.

Hello, I‘m Ben from Germany and I’m writing my bachelor thesis about redirected walking. We have pupil labs for the vive and vive pro but for my studies we need a Vive Cosmos. My question is, if it’s possible to use a pupil labs device with the Vive Cosmos. I haven‘t found any information about it. I hope you can help me, thank you!

@user-ba3796 from what I have heard, the vive addon also fits the vive cosmos. I would give it a try!

Oh, that sounds good! Thank you for the fast answer!

Okay, I'll try here.

I'm looking for a more-accessible explanation of what the "confidence" ratio is in Pupil outputs. Here's what it says in Kassner, Patera, & Bulling, 2014:

Detect edges using Canny [14] to find contours in eye image. Filter edges based on neighboring pixel intensity. Look for darker areas (blue region). Dark is specified using a user set offset of the lowest spike in the histogram of pixel intensities in the eye image. Filter remaining edges to exclude those stemming from spectral reflections (yellow region). Remaining edges are extracted into into contours using connected components [29]. Contours are filtered and split into sub- contours based on criteria of curvature continuity. Candi- date pupil ellipses are formed using ellipse fitting [16] onto a subset of the contours looking for good fits in a least square sense, major radii within a user defined range, and a few additional criteria. An augmented combinatorial search looks for contours that can be added as support to the candidate ellipses. The results are evaluated based on the ellipse fit of the supporting edges and the ratio of supporting edge length and ellipse circumference (using Ramanujans second approximation [18]). We call this ratio “confidence”.

Can anyone break this down for me (and my reviewers, I'm sure)? The ratio between what and what?

Are there any other papers I could look at? Any technical documents?

@user-36b0a3 In the 2d detection and mapping mode: Think of it as the number of edge pixels of the pupil contour divided by the analytical circumference of the pupil defined as an ellipse. 1 means the entire pupil edge was visible. .5 means half is visible. In the 3d case is gets more complicated as we take a weighted avg of the 2d confidence and 3d model agreement with the 2d result.

@mpk Thank you!!!

Okay, so I figured it must be some kind of residual. What you're saying, if I understand it correctly, is that it's the ratio of observed pixels of the circumference of the pupil, as defined by the line between the dark of the pupil and the edge of the iris according to the visual recognition algorithm for dark pupil detection, AND the expected circumference as predicted by the modeled ellipse. Is that an acceptable/reasonable explanation?

Second question: Is the recommendation of staying above ~.60 based on any theoretical premise?

Followup to that: Out of all my data, I never see it go below that. As in, the lowest value I ever see on any one person is like .60000001 or something. This makes me worry that it can't actually go lower and that that is somehow effectively a floor value...