Digitize Module

Hello Nancy,

I will look into the possibility of any facilities that are close to you so
you could see the APAS in action. I know that we have systems at
University of Missouri-Columbia (older APAS) and also at Washington
University Medical School-St. Louis but I do not know if they are willing
to provide any training or demonstrations.

In general, the 2D analysis should be fairly straight forward. I will
provide a brief description below. Also remember that each software module
has alot of information in the associated help screens!

Filming
=======
You should record a "calibration fixture" with a minimum of 4 coplanar
points that will encompass the plane of the activity that you intend to
analyze. For example, the four corners of a door frame might work. Then
record the activity of interest as it takes place in this same plane.

CAPTURE MODULE
==============
Using one of the video cameras, or any VCR, connect the video output to the
video input of the Iomega Buz frame grabber and open the Capture module.
Give the file a name and set the Capture Parameters. Then play the video
tape and select the GO icon when you are ready to catpure the desired
portion of the video to an AVI file. Each camera view should have two AVI
files: one for the calibration fixture and another for the sequence to be
analyzed.

TRIM Module
============
Since the computer captures video to an AVI file in real-time, you probably
captured much more data than is required for the analysis. The TRIM module
is used to "clip-out" the desired portion that will be used for your study.

DIGITIZE Module
===============
Select FILE, SEQUENE and NEW to begin a new sequence file for analysis and
specify the requested information. Select FILE and NEW VIEW, specify the
information and retrieve the desired AVI file for digitizing. Then select
CONTROL, DIGITIZE to retrieve and digitize the control fixture. Select
CONTROL and FINNISH and begin digitizing the AVI data file.

TRANSFORM Module
================
Open up the Transform module and select the desired data file to transform
into image space coordinates. The software will automatically detect that
it is a 2D file based on the calibration points. Select the 2D icon to
complete the transformation.

FILTER Module
==============
Open the Filter module and select the desired file to perform filtering of
the data. The filter module is used to remove "random digitizing" error.
Proceed through each joint for the smoothing.

DISPLAY Module
==============
The Display module is used to presentation of the results. Select the NEW
icon and 3D icon. Then select NEW 3D button to specify the parameters to
display/graph. You can also disply the stick figure and numerical data by
using the desired icons.

I hope this information is helpful. I recommend that you also refer to the
Quick Reference section of the associated help screens. This provides a
step-by-step approach to the basic APAS functions. If you require a more
detailed explaination, you can find this in the help screens also.

Another option would be to perform the on-line tutorials from the Ariel
internet site. The direct address is:

http://24.10.158.215/topics/Tutorials/default.htm

Please contact me for any additional information.

Sincerely,

John Probe

At 01:48 PM 11/11/1999 -0600, you wrote:
>John,
>
>We are trying to use Ariel for 2D applications, and have run into a spot of
>trouble. It occured to me that our best bet for learning the system might
>be to watch someone who already uses it. Could you tell us if any
>Universities or anyone else near the St. Louis area uses the Ariel systems?
>We could then contact them and perhaps speed up the learning curve!
>
>I'm glad Bob Proffer could help you with figuring out the payment situation.
>
>Thanks,
>
>Nancy
>
>
>
>Nancy Getchell, Ph.D.
>Division of Teaching and Learning
>University of Missouri-St. Louis
>8001 Natural Bridge Rd.
>St. Louis, MO 63121
>
>Phone: (314)516-5220
>fax: (314)516-6442
>email: Nancy_Getchell@umsl.edu
>
>

Sun-

Thank you for your list of questions. I am happy to answer them! Indeed it
is
very important to understand these variables if one is to get the most from
the autodigitizing.

> > < in Global options >
> > 1. Initial #Pix Min/Max and should be set to
When the software performs the initial locating of a spot these numbers set
limits on the size a spot can be during this initial locate phase. If the
spot is very small and one sets the min too high the software may reject an
obvious spot because it does not contain a sufficient # of pixels, and
similarly for large spots and setting the max# too low. These are outer
limits and should be set with a large margin of error. Since one points to
the image location before this initial locate operation takes place that one
set Min#=2 and max#=500. If the spots are very small you might set min#=1.

> > 2. #Pix Min/Max : There is only one input
This is the default percentage that a spot may grow or shrink by when
advancing to the next frame. For example if this is 50% and in one frame the
spot has 50 pixels, then in the next frame the spot cannot have more than
75 pixels or less than 25 pixels. This means that the software will reject
and spot that is outside these limits. If this percentage is too large the
software may find extraneous spots that are closer to the expected location.
If it is too small the software may reject an obvious spot because due to
lighting changes or motion relative to the camera has resulted is the spot
growing/shrinking by too much. This is particularly important for small
spots where a few pixels of change may be a sizable percentage.

> > 3. AutoCalc Min/Max
If this option is checked then for every frame the software dynamically
calculates a new Min/Max# pixels based on the size of a spot in the previous
frame. If it is NOT checked then the Min/Max# pixels is static, meaning that
they do not change from frame to frame but stay fixed unless explicit;y
changed.

> > 4. Relative
If this is checked then the number entered in the #Pix Min/Max is a
percentage of a spot otherwise it is a fixed number of pixels. If the is not
checked then the spot size can grow or shrink by a fixed number of pixels
frame to frame.

> > 5. Auto enhance : may be Gamma enhance?
Yes it may but within the confines of the three spot locator levels. When
"Standard" is selected no image processing is applied to the image when the
spot locating takes place. When "Enhanced" is selected the following series
of filters is applied to the image area of interest before the spot locating
takes place, Median followed by Gamma.with the Gamma factor selected using
the AutoEnhance sliderbar. On the slider the range 0 to 100 correspond to a
Gamma of 1 to 2. If "Enhanced+Edge" is selected and edge filter is added as
well. The enhancements are applied to the region of the image where the spot
is expected so one does not see the whole image transformed. However, the
before/after images shown in the Auto-Locate Properties
indeed are the filtered images.
> >
> > < in the small window poping up on the first frame of auto digi >
> > 1. Min/Max
These are the percent of spotsize [#Pixels] that will be imposed on the next
frame for this spot if relative, or absolute #pixels if not. These numbers
initially are the same as those entered in the Global Options but may be
changed point by point if desired.

> > 2. Threshold
This is the pixel brightness that a pixel must have to be considered as part
of a spot. Pixel brightness ranges from 0 to 255. The higher the threshold
the fewer pixels will be considered as part of a spot.

I hope these responses help explain how the spot locating algorithm is
controlled. It is a pleasure to explain the system.

Best, Jeremy

> ----- Original Message -----
> From: Sun G. Chung <suncg@medicine.snu.ac.kr>
> To: Jeremy Wise <Wise80x86@aol.com>
> Cc: Gideon Ariel <gideon@arielnet.com>
> Sent: Wednesday, October 27, 1999 4:19 PM
> Subject: Help!!!
>
>
> > Hi Jeremy,
> >
> > How are you?
> > I do the whole process of gait analysis from capturing to report
> generation,
> > at least one patient per week. And the auto-digi works so fine. But I
> found
> > that the auto digitization ability varies very much according to the
> setting
> > on Global Options. Sometimes it is very excellent and sometimes it is
not.
> >
> > This kind of variation seems to be increased in V4.4. So, if I should
> know
> > the meaning and effects of all the variables in Global options and if I
> > could use and handle at ease, our work will be done more easily.
> >
> > The list of things that I would like to know is
> >
> > < in Global options >
> > 1. Initial #Pix Min/Max
> > 2. #Pix Min/Max : There is only one input
> > 3. AutoCalc Min/Max
> > 4. Relative
> > 5. Auto enhance : may be Gamma enhance?
> >
> > < in the small window poping up on the first frame of auto digi >
> > 1. Min/Max
> > 2. Threshold
> >
> > If I can understand the above variables, I will be able to maximize the
> > ability of auto digi for each patient with various digitizing condition.
> > Small or large, bright or dark skin, ... ...
> >
> > If you have no time to write them, please direct me other webpages or
> > references. But your words would be the best.
> >
> > Please, help me.
> >
> > Sun

From: Wise80x86@aol.com
Date: Thu, 4 Feb 1999 10:40:14 EST
To: m.almond@ucsm.ac.uk
Cc: gideon@arielnet.com, J.Brond@mfi.ku.dk, john@arielnet.com
Subject: APAS Digi4 problem

Gideon asked me to check into the problem you reported regarding the Digi4_32
module when switching back to normal data upon completion of digitizing
control points. Thank you for reporting the problem. It is always our policy
to fix problems as quickly as possible.

This problem has been fixed and will be available to you as soon as the next
software revision is available. You will be able to download it from the net
at that time. I suggest that you keep in contact with Gideon and he can let
you know as soon as it is available.

In the mean time there are several work arounds for this problem. The problem
occurs when a second AVI file is opened from the "Control" menu item for use
with the control point digitizing while having an AVI file open for digitizing
data. There is a hitch in switching back to the original AVI file. Any one of
the following should circumvent the problem.

1) Before you go to digitize control points, select the AVI file which
contains the image for the control points as if you were selecting the AVI for
your regular data tracing. Then after you go to digitize your control points,
use the "Select Image" submenu item to select the image containing the control
points. After you have selected "finish" and are back to tracing data you will
need to reselect in the usual way the AVI file appropriate for data.

2) Digitize your control points first. When you first start a view, before you
open and AVI for tracing your data, digitize your control points.

3) After you have finished digitizing your data, close the view, reopen it
without opening an AVI and proceed to digitize your control points. At this
point there would not be a problem continuing to digitize your data.

I hope this helps & we apologize for any inconvenience.

Sincerely,
Jeremy Wise
Dir R&D, Ariel Dynamics

>Hi Gideon
>
>I seem to be experiencing a exception fault in the digitise module.
>I've copied the info that it gave me I don't know if it will help.
>
>The problem occurs when I've finished digitising the movement and then I
>open the control frame, and digitise that (the control frame is on a
>separate avi file) then when I come to close the file down I gives me
>the exception fault
>
>DIGI4_32 caused an exception 10H in module DIGI4_32.EXE at
>0137:00414975.
>Registers:
>EAX=01242c80 CS=0137 EIP=00414975 EFLGS=00010206
>EBX=00000001 SS=013f ESP=005af614 EBP=005af648
>ECX=01242c80 DS=013f ESI=01242e70 FS=2f6f
>EDX=005af678 ES=013f EDI=01242c80 GS=0000
>Bytes at CS:EIP:
>dd 5d ec 8b 55 dc db 82 ac 02 00 00 8b 45 dc da
>Stack dump:
>013f2999 bff74277 11988628 6d2417b7 01242c80 00288670 bff72999 bff62376
>01242c64 2f6f013f 0059864a 005af828 00448c70 005af6a4 0041b8b5 005af670
>
>Hope this helps
>Matt
>

Dear Dr. Dohle,

Gideon Ariel has asked me to answer your inquiry about the opposition
movement and automatic tracking of that.

We have had the apas system here for several years and use it on a daily
basis for 3D analysis of walking. It should be perfectly OK to use the
system for tracking the thumb movement in 3 dimensions. After experimenting
a little with the type of markers and their size, it will be possible to
track the movement automatically. You will have to use at least two
videocameras and a rather small calibration cube. You may build the cube
yourself, but I guess it is easier to order one from Ariel Dynamics.

The 3D coordinates can be exported in ascii, but you may also export the
apas files directly to Matlab, we always do that. Then it is easy to perform
an FFT in Matlab.

The apas system can also perform a frequency (FFT) analysis, but I think it
will be better to use Matlab or another signal processing program.

You are welcome to visit my lab if you want to see how we use the system
here. We have been very satisfied with the system and also with the service
from the company.

Sincerely yours

Erik B. Simonsen, associate prof. Ph.D.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Erik B. Simonsen, Associate Professor, M.Sc. Ph.D.
Institute of Medical Anatomy section C.
Panum Institute. University of Copenhagen
Blegdamsvej 3., DK-2200 Copenhagen N
DENMARK
Phone: +45 35 32 72 30 (work) Fax: +45 35 32 72 17
Phone: +45 45 80 93 04 (home)
http://www.biomechanics.mai.ku.dk/ebs.htm
E-mail: E.Simonsen@mai.ku.dk
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

You do not need to select anything in the global options to activate centroid detection. The program logic has been changes so that any time both the Control & Shift keys are held down when a point is digitized with the mouse the program will search for a marker in the general location of the cursor. The only item in the global options you might want to change is the "Initial Locate Dialog". If this is checked, every time theis Control & Shift capability is used an "Initial Locate" dialog will appear giving information about the spot locating process. If unchecked, this dialog is skipped.

I would suggest checking the "weighted averages" options but it is not essential.

There is a mis-understanding here. The Weighted Average factors in the "brightness" if the pixels when locating the centroid. Bright pixels count more that not so bright ones. Onder no circumstances is the "brightest" pixel used as the location of the marker. If the pixels are UnWeighted, all pixels considered as part of the marker [above marker threshold] are considered equally in locatine the centroid.

Hi Erik,

Nice to hear from you.

Yes, the contrast of the avi is very good and we can digitize an APAS view
about 1 or 2 min. It is secret how to get the contrast. But I will tell you
for the appreciation of the book introduction. Hahaha.

We use light sources for each of the cameras. The light should be directed
as to the direction of camera shot. And we use reflective markers. You may
do same as we.

The secret is to increase the shutter speed up to 1/2000 or 1/1000. Then,
you can have very contrasting marker shape even if you capture the video in
compression rate of 50-60 kb/frame.

Hope this would be helpful.

Sun

-----?? ???-----
?? ??: Erik B. Simonsen <E.Simonsen@spam.mai.ku.dk>
?? ??: suncg@medicine.snu.ac.kr <suncg@medicine.snu.ac.kr>
??: 1999? 11? 11? ??? ?? 7:47
??: markers

>Dear Sun,
>
>I am happy to hear that I have introduced you to the book of Kit V. Now I
>have a question for you. Looking at pictures from your gaitlab on Gideons
>homepage, I would VERY MUCH like to know, how you obtain such contrast for
>the white markers. I often dress the subjects in black to obtain better
>contrast, but the pictures from your lab look fantastic and easy to
digitize
>automatically. Are the pictures actually photographs or video ?????
>
>Looking very much forward to meet you.
>
>Erik
>
>++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>Erik B. Simonsen, Associate Professor, M.Sc. Ph.D.
>Institute of Medical Anatomy section C.
>Panum Institute. University of Copenhagen
>Blegdamsvej 3., DK-2200 Copenhagen N
>DENMARK
>Phone: +45 35 32 72 30 (work) Fax: +45 35 32 72 17
>Phone: +45 45 80 93 04 (home)
>http://www.biomechanics.mai.ku.dk/ebs.htm
>E-mail: E.Simonsen@mai.ku.dk
>+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Hello Mario,

Gideon has asked me to provide a better description of the digitizing
process using the PCX files. We do not normally use the PCX format for
digitizing. This was an option provided as a transition to the Windows
environment. The older APAS-DOS software could capture in either VID or
PCX formats while the APAS-WINDOWS software could capture in either PCX of
AVI. The PCX format was the "common format" between the two operating
environments. Anyways, I will list the intructions below.

NOTE:
The following description assumes the hard disk drive has been partitioned
to the C, D and E drives. The APAS program files are on C, the data (stick
figure) files are on the D drive and the video (PCX) files are on the E
drive. This also assumes that you have already captured the desired files
in PCX format.

CONFIGURING THE PXC FILE
========================
The PCX file must be configured properly so the APAS software can read
them. When images are captured in the PCX format using the "older" APAS
hardware and software this is automatically completed. However, since the
APAS was not utilized to capture, this must be performed manually.

The individual PCX pictures must reside in a directory with the same name
as the capture file. Using the files that I sent yesterday, there are 5
PCX images and the Capture file was named REACH1. Therefore, the REACH1
directory contains the following files.

E:\REACH1\REACH1.PCX
E:\REACH1\REACH2.PCX
E:\REACH1\REACH3.PCX
E:\REACH1\REACH4.PCX
E:\REACH1\REACH5.PCX

There must also be a *.PCL file on the root directory. This is just a text
file that tells the APAS software the order of the PCX images. This file
must also have the same name as the PCX directory. The E:\REACH.PCL file
contains the following information:

e:\Reach1\Reach1.pcx
e:\Reach1\Reach2.pcx
e:\Reach1\Reach3.pcx
e:\Reach1\Reach4.pcx
e:\Reach1\Reach5.pcx

This PCL file tells the APAS software the path and order of the PCX images
to be retrieved.

READY TO DIGITIZE
==================
1. Open the DIGITIZE module from the APAS SYSTEM folder.
2. Select FILE, SEQUENCE and NEW to name a new sequence.
3. Enter the Sequence Parameters (Title, Units,#Pts, # Ctrl Pts, Type,
Height, Weight, Point IDs)
4. Select FILE, NEW VIEW and enter the View Information.
5. Select FILE, OPEN PCX Images (or click the PCX icon).
6. Select the desired PCL file for digitizing.
7. Select OK at the View File Information menu.
8. The first PCX image will be displayed and you are now ready to begin
digitizing.

I have written these steps as I went through the process myself using the
current APAS Revision 4.9 software and the REACH1 PCX files. The PCX files
appeared without any problems

I hope this information is helpful. Please contact us for any additional
information.

Sincerely,

John Probe
Email: ARIEL1@ix.netcom.com

----- Original Message -----
From: Mario Lamontagne <mlamon@uottawa.ca>
To: Gideon Ariel <gideon@arielnet.com>
Sent: Wednesday, January 19, 2000 7:24 AM
Subject: RE: PCX files

> Gideon,
>
> We must be very stupid but it is not working on all the systems we have in
> the lab. I have tried on my laptop which you have installed tha APAS. What
> do you mean we are doing wrong in the setup?
> Which information do you put in the sequence?
> What frame rate do you put in the view file?
> Any Help it would be appreciated.
>
> Thanks in advance
>

++++++++++++++++++++++++++++++++++
Ariel Dynamics, Inc.
4891 Ronson Court
Suite F
San Diego, California 92111 USA
(858) 874-2547 Tel
(858) 874-2549 Fax
Email: ARIEL1@ix.netcom.com
Web Site: http://www.arielnet.com
++++++++++++++++++++++++++++++++++

-----Original Message-----
From: Ariel [mailto:ariel1@ix.netcom.com]
Sent: Tuesday, February 15, 2000 1:39 PM
To: mlamon@uottawa.ca
Cc: malha017@uottawa.ca
Subject: New Digitize Module

Hello Mario and Mouafak,

There are two options for getting the new Digitize file (that supports BMP format). You can download the latest version from the Ariel internet site, or you can simply replace the "old" file with the "new" file. As always, we strongly recommend that you make a backup (or save a copy) of the "old" file prior to any updates.

I have attached a copy of the new file. You should install it in the exact same location as the currnet DIGI4_32.EXE program and then follow the instructions as if it was a PCX file. The software now detects either PCX or BMP format.

Please let me know if you have any additional questions.

Sincerely,

John Probe
Email: ARIEL1@ix.netcom.com

========================================================
X-From_: ariel1@ix.netcom.com Tue Feb 15 11:57:28 2000
Reply-To: "ariel1 at netcom" <ariel1@ix.netcom.com>
From: "ariel1 at netcom" <ariel1@ix.netcom.com>
To: <mlamon@uottawa.ca>
Cc: "John Probe" <john@arielnet.com>,
"Jeremy - Sportsci" <jeremy@sportsci.com>
Subject: new version
Date: Tue, 15 Feb 2000 09:50:06 -0800
X-MSMail-Priority: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300

Hi Mario: The new version is on the net Gideon

Hello Susan,

I will provide answers below each of your questions.
Please contact me for any additional information.

John Probe
Email: ARIEL1@ix.netcom.com
============================

At 04:56 PM 03/31/2000 -0500, you wrote:
>I have been exploring how to use our somewhat new system and I have
>several questions/problems:
> 1. Can you graph the raw data (prior to filtering)? If so, how? If
>not, why not?
>
Yes! These procedures can be found in the Display program help file. Raw data can be graphed before or after filtering using the following steps:
a. Open the display program
b. Select the 3D icon and select the desired sequence.
c. Select the NEW 3D button to specify the data trial.
d. You will see the option to select RAW POSITION in the QUANTITY column.
e. Select the OK button to graph the data.

> 2. Can you export the raw data (before filtering)? If so, how? If
>not, why not?
>
Yes! These procedures can be found in the Display program help file. Raw data can also be exported using the following steps.
a. Follow steps a through e listed above
b. Select GRAPH, DATA to display the graph options. Make certain that RAW ONLY is selected in the Curves section.
c. Select OK to graph the data.
d. Select EXPORT, WORKSHEET, NEW and name the worksheet.
e. Select the OPEN button to create the worksheet.
f. Select EXPORT, WORKSHEET, SAVE to display the Export Channel Options menu.
g. Enter the Caption, X-Axis Start/End and X Increment values and press the OK button to save the data to the worksheet.

> 3. Why do some files for digitizing start at a time other than 0.0
>sec?
>
The 0.0 time is determined by the Synchronizing Point specified in the Digitizing Module. If no Synch Point is specified by the user, the software assumes that the first from is the Synch Point and therefore, sets the time for the first image equal to 0.0 seconds.

> 4. How can you reset the timer to start at 0.0 s?
>
The Time Value can be set to 0.0 only be specifying the Synch Point at the point where you wish the time to equal 0.0 seconds.

> 5. Why won't the export feature in 'Display' export the data points
>if the time does not start at 0.0? (I have a file that starts at 0.700
>sec going to 1.2 s. When I try to export, I get header only, no data)
>
I have tried to repeat this situation and the data exports as expected. Therefore, there is either something wrong in the procedure or something different in your data file. I suggest you follow the steps listed in Questions 1 & 2 above. If you are still experiencing problems exporting data, then you should attach the files to an Email and send them to me for examination. We would need the *.CF, *.1t (and all other related *.t files), and the *.3d file where * indicates your sequence name.

> 6. I tried to pad and manipulate the time in 'Display' and
>'Transform' but to no avail.
>
See Answers 3 & 4 above.

> 7. HEEEELLLLLLPPPPPPPP!!!!!!!!!!
>
We are always happy to assist! Hope you have a great weekend!

>Sue Chinworth
>Elon College Dept of Physical Therapy Education
>2085 Campus Box
>Elon College, NC 27244
>336-538-6861
>chinwort@elon.edu
>
>

Hello Jim,

Thank you for your Email message. If you look at the SEQUENCE information
for a project & click the SEGMENTS button, one by one all the segments
defined in the project along with the joints defining the segments, a
fractional distance between the endpoints for the segment CG, and the mass
of the segment normally expressed as a percent of the total body mass. One
can use default values or the user can specify whatever segmental model
they wish.

The center of mass for the entire body is then defined as the average of the
segmental center of masses weighted by the tfractional mass of each segment
Thus:
      Xcog=SUM(mass[i]*x[i]) / SUM(mass[i])
      Ycog=SUM(mass[i]*y[i]) / SUM(mass[i])
      Xcog=SUM(mass[i]*z[i]) / SUM(mass[i])

where the sum is over all segments and x[i],y [i], z[i] are the [x,y,z]
coordinates of the ith segment cm, and mass[i] is the mass of the ith segment.

I hope this answers your questions.

Sincerely,

John Probe
Email: ARIEL1@ix.netcom.com

At 10:53 PM 04/11/2000 -0800, you wrote:
>Dear Dr. Ariel,
>
>I am Jim from the Chinese University of Hong Kong. I would like to ask some
>question about the APAS system. As we have a "deal" before, I start to use
>the APAS-1999 system before you come to Hong Kong at June.
>
>Firstly, I would like to give a excellent comment to APAS system which
>design was easy to learn and user friendly. I only spent two days time for
>learning and working with the instruction, now I can perform the 2-D motion
>analysis under APAS-1999 system. However, I would like to solve a problem
>before I go further use of APAS-system.
>
>My problem is:
>When I creat a new sequence under the DIGI4 module, I need to creat a new
>model by selecting "Type" to "User-Defined" and then select "Segments" to
>define the segment name, segment connections, and segement mass
>information. I would like to get more instruction about the creation of
>model by "User-Defined".
>
>Sometimes, I would like to calculate the CG of the upper body only or creat
>some model for non-human study, therefore I need to creat the User-Defined
>model and to indicate the calculation method of CG. How can I get more
>information about the model design, method of CG calculation, and segment,
>joint defination method under new sequence creation ? Please let me know.
>
>I am looking forward to hearing form you. Thank you.
>
>Best regards,
>
>Jim
>
>
>
>
>
>Biomechanics Laboratory
>Department of Sports Science and Physical Education
>The Chinese University of Hong Kong
>Shatin
>Hong Kong
>Tel : (852) 2609 6079
>Fax : (852) 2603 5781
>
>

Dear John,

Thank you for your reply. I got a lot of idea on the model design.
Moreover, I have downloaded the user manual which is very helpful to me.

I Thank you for your attention.

Best regards,

Jim

At 03:02 PM 4/11/00 -0700, you wrote:
>Hello Jim,
>
>Thank you for your Email message. If you look at the SEQUENCE information
>for a project & click the SEGMENTS button, one by one all the segments
>defined in the project along with the joints defining the segments, a
>fractional distance between the endpoints for the segment CG, and the mass
>of the segment normally expressed as a percent of the total body mass. One
>can use default values or the user can specify whatever segmental model
>they wish.
>
>The center of mass for the entire body is then defined as the average of the
>segmental center of masses weighted by the tfractional mass of each segment
>Thus:
>      Xcog=SUM(mass[i]*x[i]) / SUM(mass[i])
>      Ycog=SUM(mass[i]*y[i]) / SUM(mass[i])
>      Xcog=SUM(mass[i]*z[i]) / SUM(mass[i])
>
>where the sum is over all segments and x[i],y [i], z[i] are the [x,y,z]
>coordinates of the ith segment cm, and mass[i] is the mass of the ith
segment.
>
>I hope this answers your questions.
>
>Sincerely,
>
>John Probe
>Email: ARIEL1@ix.netcom.com
>
>
>
>At 10:53 PM 04/11/2000 -0800, you wrote:
>>Dear Dr. Ariel,
>>
>>I am Jim from the Chinese University of Hong Kong. I would like to ask some
>>question about the APAS system. As we have a "deal" before, I start to use
>>the APAS-1999 system before you come to Hong Kong at June.
>>
>>Firstly, I would like to give a excellent comment to APAS system which
>>design was easy to learn and user friendly. I only spent two days time for
>>learning and working with the instruction, now I can perform the 2-D motion
>>analysis under APAS-1999 system. However, I would like to solve a problem
>>before I go further use of APAS-system.
>>
>>My problem is:
>>When I creat a new sequence under the DIGI4 module, I need to creat a new
>>model by selecting "Type" to "User-Defined" and then select "Segments" to
>>define the segment name, segment connections, and segement mass
>>information. I would like to get more instruction about the creation of
>>model by "User-Defined".
>>
>>Sometimes, I would like to calculate the CG of the upper body only or creat
>>some model for non-human study, therefore I need to creat the User-Defined
>>model and to indicate the calculation method of CG. How can I get more
>>information about the model design, method of CG calculation, and segment,
>>joint defination method under new sequence creation ? Please let me know.
>>
>>I am looking forward to hearing form you. Thank you.
>>
>>Best regards,
>>
>>Jim
>>
>>
>>
>>
>>
>>Biomechanics Laboratory
>>Department of Sports Science and Physical Education
>>The Chinese University of Hong Kong
>>Shatin
>>Hong Kong
>>Tel : (852) 2609 6079
>>Fax : (852) 2603 5781

Biomechanics Laboratory
Department of Sports Science and Physical Education
The Chinese University of Hong Kong
Shatin
Hong Kong
Tel : (852) 2609 6079
Fax : (852) 2603 5781

Hello Jim,

I will provide answers below each of your questions.

John Probe
+++++++++++++++++++

At 11:12 PM 04/12/2000 -0800, you wrote:
>Dear John,
>
>Thank you for your reply. I got a lot of idea on the model design.
>Moreover, I have downloaded the user manual which is very helpful to me.
>
>I would like to ask several questions about the User-Defined procedure
>under the SEQUENCE information.
>
When using the APAS to digitize points, the user has the option of
selecting System Defined or User Defined point types. The choice of System
option allows the specification of body joints from a predefined standard
list of names. The User-defined sequence type requires that the user enter
a name for each joint being digitized. The System option should be used
whenever human subjects are being digitized. User-defined would then be
used for non-human subjects, such as for analysis of race horces.. A
sequence may initially be defined using System joint names, then the
sequence type can be changed to User-defined to specify the names of
certain non-standard joints and segments.

>1. What is the meaning of "RadGyr"? It didn't mention in the user manual.
>
>2. There was "Abs" and "Rel" under the "Type". What does it mean ?
>
When using System-defined units, the parameter information is automatically
entered (based on Dempster's algorithm) for each segment. However, when
User-defined points are used, the APAS does not know this information and
it must be manually entered by the user. This information is entered by
selecting the SEGMENTS button in the Sequence Parameter menu.

When the SEGMENTS button is selected, the Point Connection Table will be
displayed. The top line indicated the current point. The second line
displays the connection information for the current point. Each point may
connect to as many as 5 other (lower numbered) points. For example, Point
#2 can connect to Point #1, however, Point#1 cannot connect to Point #2
because 1 is lower than 2.

In the connection information there are several parameters that must be
entered.
ConnectTo specifies the point connection where the Current point will connect
Segment is used to name the segment defined by the connection of these two
points.
Mass is distribution information that may be entered for each segment if
center of gravity and kinetic data is required for analysis.
CGFrac (Center of Gravity Fraction) is entered as a percentage of the
distance between the previous point and the current point that define the
segment. This information is always entered in Relative terms.
RadGyr is the radius of gyration for the defined segment. and is entered
in the same format and terms as the segment CG information.
Type is used to specify the Mass information as Relative or Absolute.
Relative Mass is entered as a percentage of the total body mass. Absolute
Mass is entered in weight units (kilograms or pounds).
Color is used to specify the color of the defined segment (currently not
implemented).

>3. What is the meaning of C.G. if I designed a football model which is a
>football player and the football ? Does the C.G. is the combine C.G. of the
>player and the ball ?
>
As stated above, the CG is the fractional distance between the two
endpoints for the center of gravity of the defined segment. Therefore, the
answer to this question depends on how the user defines the segments. If
the arm and ball are defined as separate segments, then each segment would
have its own segment information. If the ball is included as part of the
arm of the player, then the ball information should be added to the arm
information to calculate the segment information for the single segment.

>4. If question 3 is yes, how can I calculate the football player C.G. and
>the ball C.G. separately?
>
See answer to #3 above.

>5. What is the fractional distance between the endpoints for the segment
>CG, and the mass
>of the segment you are using? Any references?
>
When System-Defined units are specified, the APAS software automatically
uses the equations from W. Dempster 1955. Dempster's segment information
is based on the Height and Weight of the subject. This information is
entered in the Sequence Information menu. The segment information table
allows the user to enter the segment data from any desired source.

>I am looking forward to hearing form you. Thank you for your attention.
>
>Best regards,
>
>Jim
>
>
>
>
>At 03:02 PM 4/11/00 -0700, you wrote:
>>Hello Jim,
>>
>>Thank you for your Email message. If you look at the SEQUENCE information
>>for a project & click the SEGMENTS button, one by one all the segments
>>defined in the project along with the joints defining the segments, a
>>fractional distance between the endpoints for the segment CG, and the mass
>>of the segment normally expressed as a percent of the total body mass. One
>>can use default values or the user can specify whatever segmental model
>>they wish.
>>
>>The center of mass for the entire body is then defined as the average of the
>>segmental center of masses weighted by the tfractional mass of each segment
>>Thus:
>>      Xcog=SUM(mass[i]*x[i]) / SUM(mass[i])
>>      Ycog=SUM(mass[i]*y[i]) / SUM(mass[i])
>>      Xcog=SUM(mass[i]*z[i]) / SUM(mass[i])
>>
>>where the sum is over all segments and x[i],y [i], z[i] are the [x,y,z]
>>coordinates of the ith segment cm, and mass[i] is the mass of the ith
>segment.
>>
>>I hope this answers your questions.
>>
>>Sincerely,
>>
>>John Probe
>>Email: ARIEL1@ix.netcom.com
>>
>>
>>
>>At 10:53 PM 04/11/2000 -0800, you wrote:
>>>Dear Dr. Ariel,
>>>
>>>I am Jim from the Chinese University of Hong Kong. I would like to ask some
>>>question about the APAS system. As we have a "deal" before, I start to use
>>>the APAS-1999 system before you come to Hong Kong at June.
>>>
>>>Firstly, I would like to give a excellent comment to APAS system which
>>>design was easy to learn and user friendly. I only spent two days time for
>>>learning and working with the instruction, now I can perform the 2-D motion
>>>analysis under APAS-1999 system. However, I would like to solve a problem
>>>before I go further use of APAS-system.
>>>
>>>My problem is:
>>>When I creat a new sequence under the DIGI4 module, I need to creat a new
>>>model by selecting "Type" to "User-Defined" and then select "Segments" to
>>>define the segment name, segment connections, and segement mass
>>>information. I would like to get more instruction about the creation of
>>>model by "User-Defined".
>>>
>>>Sometimes, I would like to calculate the CG of the upper body only or creat
>>>some model for non-human study, therefore I need to creat the User-Defined
>>>model and to indicate the calculation method of CG. How can I get more
>>>information about the model design, method of CG calculation, and segment,
>>>joint defination method under new sequence creation ? Please let me know.
>>>
>>>I am looking forward to hearing form you. Thank you.
>>>
>>>Best regards,
>>>
>>>Jim
>>>
>>>
>>>
>>>
>>>
>>>Biomechanics Laboratory
>>>Department of Sports Science and Physical Education
>>>The Chinese University of Hong Kong
>>>Shatin
>>>Hong Kong
>>>Tel : (852) 2609 6079
>>>Fax : (852) 2603 5781
>>>
>>>
>>
>>
>
>Biomechanics Laboratory
>Department of Sports Science and Physical Education
>The Chinese University of Hong Kong
>Shatin
>Hong Kong
>Tel : (852) 2609 6079
>Fax : (852) 2603 5781
>
>

Hello Rick,

Thank you for your message. In order to obtain 3-D information, each
digitized point must be seen simultaneously by a minimum of two cameras.
These cameras should be approximately 90 degrees apart.

Each camera must also record a calibration fixture. This calibration
device can be made of almost anything as long as the X, Y, Z coordinates of
the calibration points are precisely known (relative to a single origin).

In general, the process for a 3D analysis is listed below.
1. You should capture the AVI files from each camera. Each camera should
have two AVI files; one for the sequence to be analyzed and another for the
calibration points.
2. Use the TRIM module to "clip" the desired portion of the AVI file that
will be used for the analysis.
3. Use the DIGITIZE module to digitize the sequence to be analyzed as well
as the control point information from each view.
4. Use the TRANSFORM module to transform the individual 2D images into a
single 3D image.
5. Use the FILTER module to remove "random digitizing" errors.
6. Use the DISPLAY module to present and/or analyze the results.

I recommend that you refer to the pull-down Help menu associated with each
of these programs. Each module has a section named QUICK REFERENCE that
lists step-by-step directions for a basic analysis.

Please feel free to contact us for any additional information.

Sincerely,

John Probe
Email: ARIEL1@ix.netcom.com

----- Original Message -----
From: <essner@greenapple.com>
To: <gideon@arielnet.com>
Sent: Wednesday, August 02, 2000 1:38 PM
Subject: APAS

> Gideon,
>
> I'm currently learning how to use the APAS and have run into an obstacle.
I
> would like to calculate 3-D coordinates from two camera views. How
exactly do
> you determine the values for the control point coordinates in DIGI4?
Also, I
> would like to calculate angles from the 3-d coordinates. Will APAS do
this
> calculation? How can I view the 3-D coordinates after they've been
> transformed?
>
> Thanks!
> -Rick
>
> Rick Essner
> Department of Biological Sciences
> Ohio University
> Athens, OH 45701
> (740)593-9510
> essner@greenapple.com
>
>
>
>

Hello Rick,

The X,Y,Z coordinates would have to be precisely measured for each point
(relative to a single origin and following the right-hand-rule
orientation). Once these numerical values are know, they can be entered in
the APAS software using the Digitizing module. This is performed by
following the steps listed below.

1. Open the Digitizing module
2. Select FILE, SEQUENCE, NEW to create a new sequence file. Name the
file and select the OPEN button to proceed.
3. You will see the Enter Sequence Parameters menu. Enter the Title,
Units of measure, #Points, # Control Points (8 in your case), Type of
points, Height and Weight are optional and only required for kinetic
measures. It is usually a good idea just to enter a non-zero number rather
than leave 0.
4. Select POINT IDS button to define each of the #Points entered in step 3
above. NOTE: This option is only available when System Defined points are
used.
5. Select SEGMENTS button to make segment connections between points.
NOTE: This is automatically done when using SYSTEM Type Points.
6. Select CONTROL XYZs button to enter the coordinates for each of the
calibration points. If you are using the 8 corners of the 6cm cube, then
your coordinates would look something like this:

Point X Y Z
1 0 0 0
2 6 0 0
3 6 0 6
4 0 0 6
5 0 6 0
6 6 6 0
7 6 6 6
8 0 6 6

7. Select FILE, NEW VIEW and enter the View Information for the first view.
8. Select FILE, OPEN AVI to open the AVI file for the first view.
9. Select FILE, NEW VIEW and enter the View Information for the second view.
10. Select FILE, OPEN AVI to open the AVI file for the second view.

At this point, you should have two AVI files open and ready for digitizing.
I usually recommend that the user get into a habit and digitize the
Control points first. The steps for this are listed below.

11. Click on the first view window to make it the active window.
12. Select CONTROL, DIGITIZE to let the software know you would like to
digitize the control information.
13. Select CONTROL, OPEN VIDEO, AVI and open the AVI file with the VIEW 1
Calibration fixture.
14. Digitize the Calibration points in the same order entered in step #6.

NOTE: Each image will require that you digitize a "Fixed" point as the
first point. This point can be anything visible in the video that does not
move and will not be obstructed by the movement being analyzed. This must
be the same point within each view but does not have to be the same between
views. For example, the fixed point must be the same point in the View 1
calibration and data files, but another point could be used for view 2.

15. After digitizing the calibration points for view #1, click on the View
2 windows to make it the active window.
16. Select CONTROL, OPEN VIDEO, AVI and open the AVI file for the second
view.
17. Digitize the Calibration points for view #2.
18. When finished, select CONTROL, FINISH and the display will be refreshed
with the first view from each camera. You can now proceed with the
digitizing process for the data.
19. When each image is digitized, both views can be "locked" together for
advancing by selecting the IMAGES, LOCK command. The current status of the
LOCK command will be displayed in the lower right corner of the Digitize
window.

I hope this information is a little more helpful.

Sincerely,

John Probe
Email: ARIEL1@ix.netcom.com

At 08:40 PM 08/03/2000 GMT, you wrote:
>Thanks for the help. I'm still not sure how I determine the x,y,z
coordinates
>for the calibration points. I planned on using a cube that has 8
intersecting
>points, with 6 cm sides. In the program I was using previously, you
digitized
>the cube in two views and entered a distance scale. How do I get these
>coordinates in APAS? I haven't been able to find this information in the
APAS
>manuals.
>
>-Rick
>
>Rick Essner
>Department of Biological Sciences
>Ohio University
>Athens, OH 45701
>(740)593-9510
>essner@greenapple.com
>
>> Hello Rick,
>>
>> Thank you for your message. In order to obtain 3-D information, each
>> digitized point must be seen simultaneously by a minimum of two cameras.
>> These cameras should be approximately 90 degrees apart.
>>
>> Each camera must also record a calibration fixture. This calibration
>> device can be made of almost anything as long as the X, Y, Z coordinates of
>> the calibration points are precisely known (relative to a single origin).
>>
>> In general, the process for a 3D analysis is listed below.
>> 1. You should capture the AVI files from each camera. Each camera should
>> have two AVI files; one for the sequence to be analyzed and another for the
>> calibration points.
>> 2. Use the TRIM module to "clip" the desired portion of the AVI file that
>> will be used for the analysis.
>> 3. Use the DIGITIZE module to digitize the sequence to be analyzed as well
>> as the control point information from each view.
>> 4. Use the TRANSFORM module to transform the individual 2D images into a
>> single 3D image.
>> 5. Use the FILTER module to remove "random digitizing" errors.
>> 6. Use the DISPLAY module to present and/or analyze the results.
>>
>> I recommend that you refer to the pull-down Help menu associated with each
>> of these programs. Each module has a section named QUICK REFERENCE that
>> lists step-by-step directions for a basic analysis.
>>
>> Please feel free to contact us for any additional information.
>>
>> Sincerely,
>>
>> John Probe
>> Email: ARIEL1@ix.netcom.com
>>
>>
>>
>>
>>
>> ----- Original Message -----
>> From: <essner@greenapple.com>
>> To: <gideon@arielnet.com>
>> Sent: Wednesday, August 02, 2000 1:38 PM
>> Subject: APAS
>>
>>
>> > Gideon,
>> >
>> > I'm currently learning how to use the APAS and have run into an obstacle.
>> I
>> > would like to calculate 3-D coordinates from two camera views. How
>> exactly do
>> > you determine the values for the control point coordinates in DIGI4?
>> Also, I
>> > would like to calculate angles from the 3-d coordinates. Will APAS do
>> this
>> > calculation? How can I view the 3-D coordinates after they've been
>> > transformed?
>> >
>> > Thanks!
>> > -Rick
>> >
>> > Rick Essner
>> > Department of Biological Sciences
>> > Ohio University
>> > Athens, OH 45701
>> > (740)593-9510
>> > essner@greenapple.com

Hello Slobodan Jaric,

Thank you for your message. I am glad to hear that you are making progress
with the APAS system.

When using the Automatic Digitizing option, there are several methods I can
think of to handle the situation you described.

1. Select AUTOMATIC, GLOBAL OPTIONS to display the Auto-Digitize
parameters. In the section labeled General Options, you will see an option
for Auto-Advance. When this option is selected, the digitizing will
automatically proceed to the next image. If this option is not selected,
then the user must select the advance key to advance to the next image.
This provides the option of checking the digitizing prior to proceeding to
the next image.

2. If you are monitoring the auto-digiting as it takes place, you can
select the AUTOMATIC, SUSPEND command to temporarily stop the
autodigitizing process. This would allow you to make correction on the
current image. Then select advance or reverse to continue with the
autodigitizing.

3. Select the AUTOMATIC, STOP command to stop the autodigitizing process.
When this command is selected, the software assumes that you have ended the
autodigitizing, therefore, if you wish to start again, you must select
AUTOMATIC, START and digitize each of the points again.

I hope this information is helpful. Please contact us for any additional
questions.

Sincerely,

John Probe
Email: ARIEL1@ix.netcom.com

At 09:37 AM 08/03/2000 +0200, you wrote:
>Dear Dr. Ariel,
>I guess we mainly solved our problem proper markers using table tenis
>balls, as well as higher illumination. We collected our first experimental
>file today and we had very few problems within the total number of 10,000
>frames recorded by two cameras.
>
>However, one of problems we could not solve: switching from automatic to
>manual tracking, and back to automatic. Namely, what we need is the
>following: when authomatic tracking switches to a wrong marker in a
>particular frame, to correct it manually and, thereafter to continue
>automatic tracking from the following frame.
>
>Thank you very much in advance.
>
>Regards
>****************************************
>Slobodan Jaric
>Centre for Musculo-Skeletal Research
>National Institute for Working Life
>Box.7654
>S-907 13 Umea
>Sweden
>
>Tel: /46-90-176121
>Fax: /46-90-176116
>****************************************
>

Hello Rebecca,

I had a chance to look over your data. It is hard to tell exactly without
the video, however, it appears that the 4 control points are in a different
plane than the activity being analyzed. The 4 control points seem like
they might be on a treadmill while the activity is taking place above the
calibrated area. Is this correct? If so, this could account for the
irratic results.

For 2-D analysis, a minimum of 4 control (or calibration) points must be
used. Ideally, these four points should encompass the area of the activity
to be analyzed and also lie in the same plane of the activity. The Z
coordinate must be equal to zero so the activity should take place in the
XY plane.

Is there any method to use 4 or more points to make a "calibration box"
around the turkey leg?

It appears that you are performing the correct procedures but just need to
rearrange the calibration points. Would it be possible to also send a
small video file (maybe 3 to 5 images).

Also, do not forget to contact Dr. Ariel in Brisbane. You can Email him at
the address listed below to find his location or schedule a meeting with him.

I look forward to your reply.

Sincerely,

John Probe / Dr. Gideon Ariel
Email: ARIEL1@ix.netcom.com

At 04:52 PM 9/5/00, you wrote:
>Hi John,
>
>OK - I got the additional options up in the display module after I filtered
>the data - I had thought that the filtering done in the transformation
>module was enough.
>
>Anyway, I still can't make any sense of the data once I'm into the display
>section.
>
>I have attached new files.
>
>My problem lies at this point:
>I open the 3D file in display, then I have been choosing Joint Angles,
>displacement, and then I don't understand what to do - I chose distal toe
>and prox toe, prox toe and TMT, and TMT and TT, in the hope that it would
>give me the JA info for digits, IT and ankle respectively. Also chose the
>3D rather than x,y,z because I need the whole point.
>
>But then the thing that is graphed doesn't make any sense to me, nor do the
>headings on the tabulation that shows the data points that the graphs are
>made from.
>
>All I need to obtain is the joint angles from the digitised data, both as a
>stick figure and as raw data.
>
>I hope that you can help!
>
>Is there an Ariel display at the Pre-Olympics conference that is happening
>here in Brisbane this weekend? Would there be someone there that I could
>talk to about this???
>
>Thanks again,
>
>Rebecca.
>
>Attachment Converted: "F:\MAIL\ATTUCH~1\t43sp4.1t"
>
>Attachment Converted: "F:\MAIL\ATTUCH~1\t43sp4.3d"
>
>Attachment Converted: "F:\MAIL\ATTUCH~1\t43sp4.cf"
>
>
>
>
>
>***************************************************
>Rebecca Campbell
>Department of Anatomical Sciences
>The University of Queensland
>QLD
>AUSTRALIA 4072
>
>Ph + 61 7 3365 2961
>Fax + 61 7 3365 1299
>
>Email: Rebecca.Campbell@mailbox.uq.edu.au
>
>**************************************************

Hello Eric,

The DIGITZE module can display and digitize up to 4 views simultaneously.

Synchronizing can be performed in numerous ways. First, one can have a
"synchronizing event" in the field of view. This can be as simple as a 35
mm camera flash, a foot making contact with the ground, or even a falling
object. If you desire to spend more money on a gen-lock camera system,
that APAS will fully support that also, but then you lose the advantage of
portability. Cameras can also be synchronized using using a "software
genlock" algorithm that is integrated in the TRANSFORM module. You can
access a full description of this algorithm by opening the TRANFORM module
and select HELP, INDEX and SYNCHRONIZING.

Reliability and validity information is available from the Ariel internet
site. The exact address to one such study is:

/topics/comparison/default.htm

You can also view the Bibliography section that provides a selection of
published articles using the APAS.

/Main/adw-86.html

I hope this answers your questions. Please contact us for any additional
information.

Sincerely,

John Probe
Email: ARIEL1@ix.netcom.com

At 01:15 PM 12/1/00 +0800, you wrote:
>Hello John,
>
>Thanks for your kind help. And I know how to use APAS system now.
>
>I think the APAS system is a nice software, but it can be better if it can
display 1-4 cameras at the same time and capture them synchronously to 1-4
AVI files. It's a big problem if the images from two cameras are not
synchronously. I think the APAS can provide a more convenient process to
make images synchronously, like the mention above.
>
>Finally, could you tell me some information about the reliability and
validity? and does someone use the APAS system in his published paper?
Could you tell me the title of the paper? It's very important for me if I
want to use it in my study.
>
>Eric Cheng
>
>
>

Digitizing

[Back to FAQ]