Searching \ for '[EE]: Video Camera as head position indicator - id' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/displays.htm?key=video
Search entire site for: 'Video Camera as head position indicator - id'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: Video Camera as head position indicator - id'
2001\08\10@083935 by Russell McMahon

picon face
I have been asked to investigate providing a disabled user with the means to
operate an existing book-page turner.
This device usually uses a suck / blow switch with two levels (hard / soft)
of suck and blow providing 4 commands.
The user has Multiple Sclerosis and is unable to reliably use the suck/blow
system
Other users would also have similar needs.
A complication is that minimal funding is available. This is not necessarily
a limitation but a lower cost solution would be nice.

The user has reasonable head position control and it was suggested that I
could provide a light beam pointing system where the user pointed a head
mounted light beam at 4 optical receptors mounted above the page turner.
This should be a reasonably straight forward solution:
- Use a laser pointer
- Possibly modified to provide beam modulation.although this should not bee
necessary at such short range.
- Optical saturation from ambient light should be easy to avoid in this
application.
- Provision of a timed "beam on" period with turn on enabled by eg looking
upwards at a greater than usual angle would remove the otherwise annoying
red spot on the reading material :-)

HOWEVER

It occurred to me that the availability of low cost video cameras in various
forms (serial port, parallel, USB, separate video ...) should allow a
reasonably low cost solution where the camera monitored a target and
detected head orientation. This could still include a head mounted target or
source but this could be more discrete and if active it could be infrared
and pulsed. While this sounds eminently achievable there would almost
certainly prove to be "more to it" than meets the eye at this stage. Such a
system is notionally more expensive in material cost than a headpointer /
beam system but would be more flexible, more easily adapted to related uses
and less user intrusive.
A rudimentary camera based defined-object location system could have a range
of other uses in disability aid applications.

On reflection (no pun intended) it might be easier to use the camera to
detect where a beam falls on a target than to look at the user and attempt
to determine where the beam is pointing. The "target" could be a wall or
screen above the page turner. The user need not necessarily be able to see
the beam. While this is effectively just replacing optical detectors with a
single wide area detector the above points re flexibility etc still apply.

Has anyone had experience of similar systems and / or can offer useful
suggestions that may help reduce design time.



regards

       Russell McMahon

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\10@090118 by Alan B. Pearce

face picon face
>It occurred to me that the availability of low cost video cameras in
various
>forms (serial port, parallel, USB, separate video ...) should allow a
>reasonably low cost solution where the camera monitored a target and
>detected head orientation. This could still include a head mounted target
or
>source but this could be more discrete and if active it could be infrared
>and pulsed. While this sounds eminently achievable there would almost
>certainly prove to be "more to it" than meets the eye at this stage. Such a
>system is notionally more expensive in material cost than a headpointer /
>beam system but would be more flexible, more easily adapted to related uses
>and less user intrusive.

Using a webcam type device looking over the top of the book at the person,
how would the following work out?

If the person is dark haired this may be easier, but with using IR
sensitivity could possibly work with anyone - sense the area and position of
the face e.g. look towards ceiling (or straight at camera) as you suggested
to "start command" and then rock head to one side or the other to turn page
in same direction as head rock. Sense head position by light area of face.

If the person has reasonable limb mobility it may be possible to provide
additional commands by moving left or right arm after using head movement
"start command". Sense it is an arm by long thin object instead of round
blob object you would use for face. By seeing if there is a line on the left
or right side of the image you may get more commands. If you can distinguish
between low lifted arm and high lifted arm you could have two commands per
arm.

Doing it this way the person would not need to have any headgear at all, let
alone be connected to anything.

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\10@093310 by Stephen Webb

flavicon
face
> On reflection (no pun intended) it might be easier to use the camera to
> detect where a beam falls on a target than to look at the user and attempt
> to determine where the beam is pointing. The "target" could be a wall or

Most definitely.  I believe there has been much research directed at
determining a persons gaze from camera images, and it turns out to be a
hard problem.  Even with an active piece (laser pointer) in the equation,
I think that calibration would be a nightmare.

But, detecting where the laser hits in the world would be easier, if the
problem is constrained enough (which I believe it is).

> screen above the page turner. The user need not necessarily be able to see
> the beam. While this is effectively just replacing optical detectors with a
> single wide area detector the above points re flexibility etc still apply.
>
> Has anyone had experience of similar systems and / or can offer useful
> suggestions that may help reduce design time.

My initial reaction is to use the 4 sensor + light beam
approach.  Simplicity, ease of use / configuration, and low cost are all
there.

If you do go with a camera based solution, a few things to consider:

1.  Detecting where the laser dot (or infrared light beam, or whatever) is
in the image is pretty straightforward.  You might have the best success
buying a Matrox meteor image capture/image processing board, and using the
MIL (matrox imaging library) which has fast and robust image processing
features (hardware accelerated).  This will probably make your life quite
a bit easier, but adds $$$.  I have some image processing code that you
may be interested in, contact me off list if you want it.

2.  Determining what the world point of a given image point is somewhat
more difficult unless constrained (which I believe you can do in this
case).  If you constrain the camera to be viewing an approximately planar
surface (say a desk + the book) you can compute image to world
correspondences with a planar homography.  You need 4 matchpoints (point
the laser at known world points (in the plane) and figure out where they
hit in the image.  Of course you can constrain the problem even further
and place the camera orhtogonal to the the plane, and only have to deal
with an affine transformation (scale/ rotation/translation, no skew)
(and you can get rid of the rotation part pretty easily too)

See "Three-Dimensional Computer Vision"  by Olivier Faugeras for details.


3. If determining what part of the plane (on the desk, for now) the spot
hits is not be enoguh?  (you mentioned targets, and pointing the laser at
different targets to cause different actions, etc)  Automatically
detecting different targets, (without a manual calibration phase) will
probably turn out to be fairly difficult.  Changing lighting conditions,
camera positions, etc may cause problems.   The more targets you must be
able to distinguish, the harder it becomes.  You can vary the color,
structure, and relative positioning of the targets, but it will definitely
be a task to get the system to distinguish them robustly.


This might not be the best paper, but you might find some useful info:

http://www.cs.cmu.edu/~rahuls/Research/Projector/


Good luck.

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\10@095817 by Thomas McGahee

flavicon
face
Russell,

Prepare a series of sensors, one for each desired choice.
Each sensor consists of:

 1) a container or isolation shell that is at least 1" deep.
    the purpose of the isolation shell is to isolate each
    sensor so that it only "sees" the laser beam when it
    is aimed directly at that target.

 2) The front of each sensor shell consists of a translucent
    window. Something like translucent drafting material
    works well. When light from the laser pointer hits
    the target window, the translucent target will diffuse
    the light to some extent. The target window should be
    large enough to allow the user to easily keep the laser beam
    on the target for about a second.

 3) a convex lens positioned such that it gathers light from
    the entire window area and focuses it onto a photo detector
    sensitive to the laser beam.

 4) an optional narrow band optical filter that passes only
    the laser beam. This enhances the sensitivity of the device
    to the laser beam and prevents false detections. The filter
    can be quite small if it is placed just in front of
    the sensor.

 5) an optional red LED positioned such that light from the LED
    falls on the sensor, but does not obscure the light from the
    target window. A means must be provided for adjusting the
    intensity of the LED. This LED light is used to "bias" the
    sensor into its active range. It improves sensitivity.

 6) a sensor that is good at detecting the laser beam. You may
    have to experiment around to determine which sensors work
    best.

There are several ways of using these detector modules.

METHOD 1
 Have the sensor incorporated in a circuit that produces an
 output voltage that produces two states. One state indicates
 that the light level is significantly above ambient, and the
 other state indicates that the light level is at or below
 ambient. An op-amp or comparator and a potentiometer are
 all that you really need.

 A selection is made by keeping the beam on the target
 for a short period of time.

METHOD 2
 Using an electrolytic capacitor and a resistor, monitor the
 CHANGES in output voltage. Note that there will be changes
 when lights are turned on/off, or when sunlight through a
 window changes. Avoid having direct sunlight hit the sensor
 array.

 To make a selection with this setup, move the beam onto and
 off of the desired sensor a couple of times. Software looks
 at the changes and declares a hit when there are several
 rapid changes.

Method 1 is more intuitive for the user. Visual feedback of beam
detection and choice acceptance can be indicated with two LEDs
at each sensor. Red indicates beam is detected, and green indicates
that the choice has been accepted.

Fr. Tom McGahee


{Original Message removed}

2001\08\10@100239 by goflo

flavicon
face
A lot of effort has been put into what are called "heads-up"
displays for aircrew.

I've heard of systems which track the gaze, initiate selected
action on detecting a sort of squint, which sounds like a pos-
sibility here. Perhaps a wink detector?

regards, Jack

Russell McMahon wrote:
>
> I have been asked to investigate providing a disabled user with the means to
> operate an existing book-page turner.

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\10@134008 by John Ferrell

flavicon
face
Going back in time here....

If you focus a video camera on a paper test pattern with regular spaced
vertical stripes
It will generate very accurate pulses when that area is scanned.  Remember
that the total scan line time is about 65 micro seconds.  This allowed us to
work on the frequency response of video amplifiers way back when.

Raster scanning has not always been the dominate method for OCR either. I
have an unreasonable amount of experience with flying spot scanners that
worked really well thirty years back.

John Ferrell
6241 Phillippi Rd
Julian NC 27283
Phone: (336)685-9606
Dixie Competition Products
NSRCA 479 AMA 4190  W8CCW
"My Competition is Not My Enemy"



{Original Message removed}

2001\08\10@145618 by Douglas Butler

flavicon
face
Might I suggest an accellerometer package such as a 3 axis Crossbow.  It
is small and light and costs about US$300.  The user could swirl his/her
head to turn pages foreware and backward.  Various nods could also be
decoded.  I think it could be very intuitive.
Check out http://www.xbow.com product CXL04M3.

Just my $.02, Sherpa Doug

> {Original Message removed}

2001\08\10@152942 by Bob Barr

picon face
Douglas Butler wrote:
>
>Might I suggest an accellerometer package such as a 3 axis Crossbow.  It
>is small and light and costs about US$300.  The user could swirl his/her
>head to turn pages foreware and backward.  Various nods could also be
>decoded.  I think it could be very intuitive.
>Check out http://www.xbow.com product CXL04M3.
>

Another non-contact alternative that might be worth considering is
ultrasonic sensing for head movements.

As I recall, a member of the Forth Interest Group (David Jaffe, I believe)
did a lot of work on this with Stanford Hospital or perhaps with one of its
clinics.

He used Polaroid ultrasonic ranging units to sense head movements which then
controlled an electric wheelchair's speed and direction.

Regards, Bob


_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\10@153258 by Stephen Webb

flavicon
face
> Might I suggest an accellerometer package such as a 3 axis Crossbow.  It
> is small and light and costs about US$300.  The user could swirl his/her
> head to turn pages foreware and backward.  Various nods could also be
> decoded.  I think it could be very intuitive.
> Check out http://www.xbow.com product CXL04M3.
>
> Just my $.02, Sherpa Doug

I like this idea.

-Steve

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\10@153739 by David VanHorn

flavicon
face
At 03:29 PM 8/10/01 -0400, Bob Barr wrote:
>Douglas Butler wrote:
>>
>>Might I suggest an accellerometer package such as a 3 axis Crossbow.  It
>>is small and light and costs about US$300.  The user could swirl his/her
>>head to turn pages foreware and backward.  Various nods could also be
>>decoded.  I think it could be very intuitive.
>>Check out http://www.xbow.com product CXL04M3.
>
>Another non-contact alternative that might be worth considering is
>ultrasonic sensing for head movements.

The head could be the center plate in a split capacitor.
Use the value differential to determine position, and can work in two axis
at least.
Almost a charge-transfer sort of thing, but not quite.
--
Dave's Engineering Page: http://www.dvanhorn.org

I would have a link to http://www.findu.com/cgi-bin/find.cgi?KC6ETE-9 here
in my signature line, but due to the inability of sysadmins at TELOCITY to
differentiate a signature line from the text of an email, I am forbidden to
have it.

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\10@154256 by goflo

flavicon
face
Douglas Butler wrote:
>
> Might I suggest an accellerometer package such as a 3 axis Crossbow.  It
> is small and light and costs about US$300.  The user could swirl his/her
> head to turn pages foreware and backward.  Various nods could also be
> decoded.  I think it could be very intuitive.
> Check out http://www.xbow.com product CXL04M3.

A neat idea - Smaller, lighter, and way cheaper would be Analog Devices
ADXL202, which has 2 axis of sensitivity and a PIC-friendly PWM output.
For the proposed use AD would likely sample out some if asked.

regards, Jack

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\10@160932 by Micro Eng

picon face
Russell

I havent read all the replies...but looke at doing a logigtech or like
system that the user has a tranducer on the head and a reciever (ultrasonic)
mounted above to determine position. Email me direct if you want more info.



_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\10@185450 by steve

flavicon
face
> If you focus a video camera on a paper test pattern with regular
> spaced vertical stripes It will generate very accurate pulses when
> that area is scanned.

This reminded me of a technique that can be used for focussing in
automated telescopes with CCD cameras. Because a star is a
point source, you can just look at the peak on a single line and
adjust the focus to get it as narrow as possible. The circuitry to do
that is really simple compared to an image processing approach.

In Russell's case, an image of a flat surface with a laser dot
through a good optical filter is going to be a similar image. Using
the same approach, it is going to be relatively simple to determine
the position by looking for a peak X microseconds accross and Y
lines down in the frame.

However, probably the easiest and cheapest solution if the person
is close to the target would be a narrow beam red and IR LED
combination on the head and IR detectors on a board. Modulate
the IR led at 38kHz and use a standard detector and the red LED
provides some visual feedback to the user. Have both the source
and detector in a bit of tube to make it even more directional.


Steve.

======================================================
Steve Baldwin                Electronic Product Design
TLA Microsystems Ltd         Microcontroller Specialists
PO Box 15-680, New Lynn      http://www.tla.co.nz
Auckland, New Zealand        ph  +64 9 820-2221
email: spam_OUTstevebTakeThisOuTspamtla.co.nz      fax +64 9 820-1929
======================================================

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\11@043706 by Peter L. Peres

picon face
Russel, maybe you could adapt one of the miniature wireless mice or
trackballs. Use the XY axes mechanically to sense head turning and tilting
(or nod & tilt) and maybe sense blinking reflectively. I see this done
using a pendulum type sensor, encased in a small tube filled with oil for
damping. I have made a reflective blink sensor using a lensless spectacle
frame and a modulated IR LED + SFH type detector. It works well. It senses
the pupil to some extent so it could be used for gaze sensing but this
would require a lot of work imho. Several commercial gaze sensors use
infrared cameras observing the pupil. Look for patents etc imho. if you
want to go with a camera type.

Peter

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2001\08\12@033741 by Dave Hatton

flavicon
face
One alternative would be to provide a sequencer, started and stopped by the
user.

Dave
----- Original Message -----
From: "Russell McMahon" <.....apptechKILLspamspam@spam@CLEAR.NET.NZ>
To: <PICLISTspamKILLspamMITVMA.MIT.EDU>
Sent: Friday, August 10, 2001 1:31 PM
Subject: [EE]: Video Camera as head position indicator - ideas sought


> I have been asked to investigate providing a disabled user with the means
to
> operate an existing book-page turner.
> This device usually uses a suck / blow switch with two levels (hard /
soft)
> of suck and blow providing 4 commands.
> The user has Multiple Sclerosis and is unable to reliably use the
suck/blow
> system
> Other users would also have similar needs.
> A complication is that minimal funding is available. This is not
necessarily
> a limitation but a lower cost solution would be nice.
>
> The user has reasonable head position control and it was suggested that I
> could provide a light beam pointing system where the user pointed a head
> mounted light beam at 4 optical receptors mounted above the page turner.
> This should be a reasonably straight forward solution:
> - Use a laser pointer
> - Possibly modified to provide beam modulation.although this should not
bee
> necessary at such short range.
> - Optical saturation from ambient light should be easy to avoid in this
> application.
> - Provision of a timed "beam on" period with turn on enabled by eg looking
> upwards at a greater than usual angle would remove the otherwise annoying
> red spot on the reading material :-)
>
> HOWEVER
>
> It occurred to me that the availability of low cost video cameras in
various
> forms (serial port, parallel, USB, separate video ...) should allow a
> reasonably low cost solution where the camera monitored a target and
> detected head orientation. This could still include a head mounted target
or
> source but this could be more discrete and if active it could be infrared
> and pulsed. While this sounds eminently achievable there would almost
> certainly prove to be "more to it" than meets the eye at this stage. Such
a
> system is notionally more expensive in material cost than a headpointer /
> beam system but would be more flexible, more easily adapted to related
uses
> and less user intrusive.
> A rudimentary camera based defined-object location system could have a
range
{Quote hidden}

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2001\08\12@180537 by den

picon face
How about a simple head mounted angle/inclinometer ? A very cheap one woul be one pot. fitted on a head band (used for forehead fitted lamps), and a small weight hanging down from the axle that is produting horizontally near an ear. Up/down movement of the head would vary the resistance. Use a high pass filter function. If possible, a second could be fitted over the back of the neck, if the person can tilt his/her head.

Use a small PIC with A/D for relevance ;=)


Sven in Sweden

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2001\08\12@203819 by Robert Rolf

picon face
Been there done that. Pain In The Azz.

The trick is to get enough damping into the pendulum so that it doesn't
oscillate. Best results came from eddy current damping using an aluminum
pendulum and strong magnets on either side of the vane. Use optical
position sensing to minimize friction. Strobing the optos gave us
very low power consumption.

We used a welders headband since that gave a nice high mount point.

Today I use the ADXL202 (only need one axis), and with a
FET switch on the filter capacitor, I can strobe it too.

R

"Milton Medicinteknik KB, Vikingstad, Sweden" wrote:
>
> How about a simple head mounted angle/inclinometer ? A very cheap one woul be one pot. fitted on a head band (used for forehead fitted lamps), and a small weight hanging down from the axle that is produting horizontally near an ear. Up/down movement of the head would vary the resistance. Use a high pass filter function.

High pass?? Then all you get is the oscillations...

If possible, a second could be fitted over the back of the neck, if the
person can tilt his/her head.
>
> Use a small PIC with A/D for relevance ;=)
>
> Sven in Sweden
>
> --
> http://www.piclist.com hint: The PICList is archived three different
> ways.  See http://www.piclist.com/#archives for details.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


More... (looser matching)
- Last day of these posts
- In 2001 , 2002 only
- Today
- New search...