Webgazer.js webcam eye tracking on the browser • Hacker News Board

Re: Webgazer.js webcam eye tracking on the browser

#11

Post by ebg13 » Thu, Jul 16, 2020, 5:40 PM UTC

lazyjeff wrote: post #10

ebg13 wrote: post #8

I tried my best on the calibration demo page in Safari and consistently get a 0% accuracy result. It calibrates sort of ok in Firefox. I guess this is heavily browser dependent? But why?

Hmm, not sure -- do you see the gaze dot when you're looking at the center circle?

I see a gaze dot. It just seems kinda like Safari is accumulating some tracking error that Firefox doesn't. Both browsers find my face accurately.

(side note: when the gaze dot overlaps the calibration circles, the circles become unclickable without some very careful fiddling)

Re: Webgazer.js webcam eye tracking on the browser

#12

Post by cbrgm » Thu, Jul 16, 2020, 6:16 PM UTC

Can you recommend any literature to get an introduction into eye-tracking at all?

Re: Webgazer.js webcam eye tracking on the browser

#13

Post by natebleker » Thu, Jul 16, 2020, 6:26 PM UTC

xamuel wrote: post #5

natebleker wrote: post #2

Disclaimer, I am the primary engineer of a commercial eye tracking system. Tools like these popup every now and then are a nice tool for rough estimation of gaze. Fixation tracking is a large complicated issue that usually requires some sort of calibration to get more precise results. By sidestepping the calibration problem, much higher subject compliance can be achieved since you don't have a grad student barking co…

In your estimation, how far away are we from eye-tracking software being able to detect the start of a microsaccade, estimate where the gaze is moving toward, and draw new contents on the screen there before the gaze even reaches that point? I would think that by "hacking" into human brain "vulnerabilities" like saccadic masking and chronostasis, such software could potentially yield seriously trippy and mind-alterin…

There’s two parts to that implementation of such a system, and they’re both interesting! The first part is detection of a micro saccade which is already available in research systems I have personally worked on. You can basically crank up the camera frame rate until you’re around the micro saccade range and do some clustering analysis on the positional data to decouple the hardware moving vs the face. The second part is having a fast enough reliable commercial grade displa system to present stimulus on. Displays are very fickle in practice and getting your hands on one that can run with adequate color, contrast, and brightness at speed is currently very difficult. One angle under research currently is the perception of stimulus during saccades and micro saccades. There’s quite a bit of time and effort in the industry going into neurological assessments through saccades, the tools coming out of this are really starting to come down in price. This opens the door for a bunch of lower priced research options, such as the parent article, to enable a much more rapid pace of understanding.

Re: Webgazer.js webcam eye tracking on the browser

#14

Post by stared » Thu, Jul 16, 2020, 6:55 PM UTC

I had an idea for an evil-eye game, in the style of Happy Tree Friends or so. That is: a lot of cute creatures around, and nothing happens (they may giggle, jump or blink occasionally) until you look at them. If you do...

Re: Webgazer.js webcam eye tracking on the browser

#15

Post by natebleker » Thu, Jul 16, 2020, 6:57 PM UTC

cbrgm wrote: post #12

Can you recommend any literature to get an introduction into eye-tracking at all?

Here’s an old school high level overview of pupil tracking methods[0]. The focus recently has been moving these older computer vision methodologies to ML without introducing too many errors.

[0] http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.155...

Re: Webgazer.js webcam eye tracking on the browser

#16

Post by Meph504 » Thu, Jul 16, 2020, 7:28 PM UTC

The listed use cases here struck me as odd, I can't imagine why I or anyone wanting to read a news article would allow the site access to my webcam. Nor do I think most users would turn on their cameras to help with any given sites analytics.

So aside from games, what is the actual use case here ?

Re: Webgazer.js webcam eye tracking on the browser

#17

Post by lazyjeff » Thu, Jul 16, 2020, 7:35 PM UTC

Meph504 wrote: post #16

The listed use cases here struck me as odd, I can't imagine why I or anyone wanting to read a news article would allow the site access to my webcam. Nor do I think most users would turn on their cameras to help with any given sites analytics. So aside from games, what is the actual use case here ?

Great question. There's a use for usability testing, where paid participants currently sit at an application in a lab to do eye tracking. So there can be a similar version online for if let's say, Bloomberg is testing a new version of their online terminal. Multiple large companies exist that solely develop eye trackers, so there's quite a market for this.

Plus it does not have to be monetary compensation. I could imagine eye tracking as an option to get around the current newspaper paywalls. If you want to read an article, you can choose to pay (subscription) as usual, or you can read them for free if you let the New York Times see what articles you read, for how long, in which order, or what ads you looked it.

Re: Webgazer.js webcam eye tracking on the browser

#18

Post by pmlnr » Thu, Jul 16, 2020, 7:38 PM UTC

Now add this to vidconf systems by figuring out which speaker is who, and make a button to talk only to them.

Re: Webgazer.js webcam eye tracking on the browser

#19

Post by awake » Thu, Jul 16, 2020, 7:40 PM UTC

In the brown undergraduate computer vision course one of the homework’s (or final project; my memory is a little hazy) was to improve the tracking of webgazer. People were pretty successful from what I remember. As an undergrad I didn’t think much of it but as a grad student I would be slightly mortified if undergraduates were improving on my current research for homework.

Re: Webgazer.js webcam eye tracking on the browser

#20

Post by jasonjayr » Thu, Jul 16, 2020, 8:05 PM UTC

Meph504 wrote: post #16

The listed use cases here struck me as odd, I can't imagine why I or anyone wanting to read a news article would allow the site access to my webcam. Nor do I think most users would turn on their cameras to help with any given sites analytics. So aside from games, what is the actual use case here ?

Nielson Research in years past would ask volunteers to attach special equipment to their televisions to report back what they were watching, and for how long. Even nowadays, I have a relative that carries an always-listening device to track ambient music being played/heard.

A technology like this would make it easier to run a program with a subset of volunteers online, w/o them needing any special equipment.

Of course, given that slightly more people (than tech communities) are aware of the dangers of an uncovered camera, the possibility of using this broadly w/o consent is nearly nil.