The EyeWriter is a low-cost eye-tracking apparatus + custom software that allows graffiti writers and artists with paralysis resulting from Amyotrophic Lateral Sclerosis to draw using only their eyes.
The original design, as shown here, featured a pair of glasses as the basis for the eyewriter design:
This eyewriter system is cheap, and completely open source. At the moment, it costs about 200$ in parts. Traditional commercial eye trackers costs between $9000-$20,000, so this is a magnitude of order cheaper, and is designed to help anyone who wants or needs an eyetracker.
This fall, we've been showing off and demoing the 2.0 device -- check out the eyewriter 2.0 in action -- we even hooked it up to a robotic arm, to draw the artwork people make with their eyes:
http://www.switched.com/2010/12/13/eyewriter-teams-up-with-robotagger-to-print-kids-ocular-artwork/print/
(The 2.0 device was designed with help and input from Takayuki Ito, Kyle McDonald, Golan Levin and students of the eyewriter collab at Parsons MFADT. Thanks also to the Studio for Creative Inquiry / CMU for hosting a session for development)
The original design, as shown here, featured a pair of glasses as the basis for the eyewriter design:
Since that first video, we've been hacking on and developing the project, and we have a new design, which we've called "eyewriter 2.0" which improves the accuracy of the device, and allow for people who's heads are moving slightly to also use an eye tracker. The original eyewriter, designed for a paralyzed Graffiti artist TEMPT1, is designed to be worn on a completely motionless head. The 2.0 design, which uses a camera and LED system mounted away from the head, can be used by people whose heads are moving slightly, such as MS patients, and people who wear glasses, etc.
This eyewriter system is cheap, and completely open source. At the moment, it costs about 200$ in parts. Traditional commercial eye trackers costs between $9000-$20,000, so this is a magnitude of order cheaper, and is designed to help anyone who wants or needs an eyetracker.
This fall, we've been showing off and demoing the 2.0 device -- check out the eyewriter 2.0 in action -- we even hooked it up to a robotic arm, to draw the artwork people make with their eyes:
http://www.switched.com/2010/12/13/eyewriter-teams-up-with-robotagger-to-print-kids-ocular-artwork/print/
(The 2.0 device was designed with help and input from Takayuki Ito, Kyle McDonald, Golan Levin and students of the eyewriter collab at Parsons MFADT. Thanks also to the Studio for Creative Inquiry / CMU for hosting a session for development)
Remove these ads by
Signing UpStep 1: Overview
The basic idea approach is that we'll be doing a few things. First, we'll be making LED illuminators for the sides of the screen and the center. Second, we'll be hacking the PS3 eye camera to get the vertical sync (when the frame of video is being taken) and to make it sensative to IR. Third, we'll be programming and building the arduino / cirucit to control the blinking. Finally, we'll setup the base for the system and go through the basics of the software.
From a technical perspective, the 2.0 system works by strobing 3 IR illuminators every frame. On even frames, it uses the center illuminator (located around the camera lens) and on odd frames it uses the 2 side illuminators. On even frames, the pupil appears bright, since the IR light is actually bouncing off the back of your eye, like red eye effect. On odd frames, your pupil appears dark. The difference between the two allows us to isolate and track the pupil in realtime. Additionally, the glints (reflections of the IR illuminators) of the dark frame are tracked, and these, plus the info on the pupil, is calibrated to screen position using a least squares fitting process for an equation that provides a mapping of glint/pupil position to screen position.
From a technical perspective, the 2.0 system works by strobing 3 IR illuminators every frame. On even frames, it uses the center illuminator (located around the camera lens) and on odd frames it uses the 2 side illuminators. On even frames, the pupil appears bright, since the IR light is actually bouncing off the back of your eye, like red eye effect. On odd frames, your pupil appears dark. The difference between the two allows us to isolate and track the pupil in realtime. Additionally, the glints (reflections of the IR illuminators) of the dark frame are tracked, and these, plus the info on the pupil, is calibrated to screen position using a least squares fitting process for an equation that provides a mapping of glint/pupil position to screen position.
Visit Our Store »
Go Pro Today »
http://www.ebay.co.uk/itm/25mm-Board-Lens-Day-Night-for-CCTV-Cameras-Foscam-Cameras-/110971149583?pt=UK_CCTV&hash=item19d6655d0f#ht_685wt_890
Thx in advance for your help
Julien
When it says to "build and run the software" does it mean to simply run the xcodeproj file or am I missing something?
Any and all help would be greatly appreciated!
When I try to run the Xcodeproj file I get the errors in the picture included
I was wondering if I can use another MOSFET other than the one specified since I can't find it in my country given these specifications. I read its data sheet and it shows that it can handle up to 36A, why do we need that much amperes anyways?
I know many of you have questions regarding the electronic parts. I am glad to help as much as I can! I can answer questions in English, French, Spanish, Swedish, Cambodian, Korean and Japanese.
Please visit my blog. I'm working on a tutorial for the eyewriter with a little bit more details, and possibly more specific explanations.
Cheers
http://hikarielectronics.blogspot.jp/
Cheers!
I am at step 11 with assembling the final breadboard which besides the programming would be my weakest point not having much electronic experience. I sourced most of the parts here in Sweden except for the Lens and lens holder.
The transistor I received is of the "older" kind (see attached image) and I am unsure of where the collector, base and emitter should go (excuse my electric ignorance ;) seeing as we are connecting it to two positives and one negative?
Seeing as I am such a novice it has been tough "reading between the lines" where I guess the author took for granted anybody attempting this would have some kind of base knowledge (which I have had to google ;) So it would be nice to hear somebody say: 'Now I have finished the whole project and it works like a charm!' Pls some light at the end of the adventurous tunnel ;)
Thanks in advance!
http://www.openframeworks.cc/download/older.html
One Question: We use a LOREX IR Baby Monitor to keep tabs on her when we are not in the room with her. We have a spare camera unit, which has the IR LED's already integrated. I'm not sure if it strobes the IR's, or how to find the vSync pin. Can anyone recommend someplace I can find this information, or if anyone has had success using this type of camera before?
Thanks,
CK
I work at a science museum and we are trying to get this working. The guys in the shop have done their part with all the hardware, and now it's up to me, for the software side. I've gotten the software running on both Windows 7 and Mac OS X, and it does recognize my eye-- however, I can't seem to get to the different modes (calibration, drawing, pong, etc) from the app. If I change the mode in the code and re-compile, I can see those different screens. But it sounds like I should be able to press the spacebar from within the main app, and find the calibration screen. Seems like this is crucial to get to interact with it as a mouse!
I'd really appreciate any pointers or suggestions you might have.
http://cid-f966c677a7c86219.office.live.com/embedicon.aspx/Remote%20Eye%20Tracker%20Windows
Refer to the instructables to change from demo mode to live. No documentation on the settings so you are on your own. Took me a long time to find the sweet spot to get the glints to lock on, but I overlook a lot of obvious things. Couldn't get the typing program to speak the words but I'm thrilled I got it to work at all.
Real
Uploaded the cbp file and others in the directory, then zipped everything except the bin directory which is already zipped and uploaded. Lemme know if thats what you needed.
Thanks!
OF_ERROR: Error allocating a video device.
OF_ERROR: Please check your camera with AMCAP or other software
AMCAP and the CL Eye Test both show that the camera works correctly. Not sure what to do next. Tried values 0,1,2,3,4 under DeviceId in the settings file but it still shows "SETUP: Setting up device 0"
I had a little luck emailing the people who wrote the code directly when I got stuck compiling but the question had to be very specific to get a usable answer. Most responses were "I don't know what you are doing wrong" and "Keep trying until you figure it out".
Thanks!
Thank you!!
https://skydrive.live.com/?cid=f966c677a7c86219&sc=documents&id=F966C677A7C86219%21103
I recently started to build an eyetracker and tried the eyewriter.
First question:
I use the linked eyewriter project, but it looks as the first eyewriter-version and calibration does not work... The software UI looks much simpler. (I got the eyewriter code working with the current open frameworks version...)
Second question:
Do I really have to do the strobing stuff? I want to build an head mounted device, using right now just one IR-Led... does the software work with a head mounted device, using only one, not strobing ir-led?
I am not a programmer. Trying to compile the project to run in Win XP via Code::Blocks and after shuffling missing files into place, I'm now at line 52 of trakingmanager_controlpanel.cpp with error message ofxLibdcPtGrey was not declared in this scope & cam was not declared in this scope. Where should I go for help?
Better yet (as russ_hensel posted), is there a windows executable of the eyewriter software available anywhere?
Thank you for posting the project!
Real