Tweet
Tochi:
> i believe that extrapolation techniques are used for
> missing surface marker data, for a certain number of trials
> [five?]
I am not sure that I understand what you mean in this context by
"extrapolation". I presume that you mean in most cases "INTERpolation",
where the digitized values for missing frames are reconstructed by fitting
curves to the available data, and then finding interpolated values for the
unavailable times. Yes, that would be OK if you are only missing a frame or
two. But my concern is with what happens when, from the viewpoint of one of
the cameras, a marker goes to the "other" side of the leg (for instance),
and stays there for the rest of the trial. We would not be able to solve
such a problem using interpolation.
From what you say, I assume also that you are looking for data that
are averages for a number of trials (say, five). In an activity like gait
(whether it is walking or running), this may be OK. But in more explosive
activities (lika a high jump, for example), most often we would be
interested in what happens in each single trial. This is because, even
though the technique of an individual is to a certain extent like a
fingerprint, similar every time, there ARE some differences between one
trial and the next, and we are interested in knowing what those differences
are.
> if the subject is involved in a sport that requires some
> twisting, somersaulting, etc in conjunction with
> translation, ie motion in more than one plane and quadrant,
> what happens? will alll segments of interest be actually
> filmed appropriately for analysis?
No, we will not be able to see all landmarks throughout the entire
activity (unless we use more than two cameras, which is possible, but in
practice presents problems of logistics, extra cost, and extra effort on the
part of the researchers). In the frames in which a landmark is hidden (for
instance, if we are digitizing the elbow joint, and it goes behind the trunk
for a few frames), we estimate its approximate position on the film as
closely as we can. But I think that such an estimation would be harder to
do when a SURFACE marker becomes hidden.
> would it not be better to get the same subject to perform
> under controlled conditions [eg in a laboratory], and thus
> use surface markers and the associated equipment?
The greatest problem would be that athletes often are not able to
repeat in non-competitive conditions the results that they achieve in
competitions. For instance, when Dick Fosbury won the Olympics in 1968 with
a 2.24 m high jump, the best that he had ever been able to clear in training
was 2.13 m.
> in situations like pathological gait analyses [eg for the
> elderly and infirm], i think that surface markers would be
> oe of the major ways to go. disagree? ;-)
I think you are probably right, in particular because in those
activities there will not be complex twisting movements to hide different
landmarks in different parts of the trial.
---
Jesus Dapena
Department of Kinesiology
Indiana University
Bloomington, IN 47405, USA
1-812-855-8407 (office phone)
dapena@valeri.hper.indiana.edu (email)
> i believe that extrapolation techniques are used for
> missing surface marker data, for a certain number of trials
> [five?]
I am not sure that I understand what you mean in this context by
"extrapolation". I presume that you mean in most cases "INTERpolation",
where the digitized values for missing frames are reconstructed by fitting
curves to the available data, and then finding interpolated values for the
unavailable times. Yes, that would be OK if you are only missing a frame or
two. But my concern is with what happens when, from the viewpoint of one of
the cameras, a marker goes to the "other" side of the leg (for instance),
and stays there for the rest of the trial. We would not be able to solve
such a problem using interpolation.
From what you say, I assume also that you are looking for data that
are averages for a number of trials (say, five). In an activity like gait
(whether it is walking or running), this may be OK. But in more explosive
activities (lika a high jump, for example), most often we would be
interested in what happens in each single trial. This is because, even
though the technique of an individual is to a certain extent like a
fingerprint, similar every time, there ARE some differences between one
trial and the next, and we are interested in knowing what those differences
are.
> if the subject is involved in a sport that requires some
> twisting, somersaulting, etc in conjunction with
> translation, ie motion in more than one plane and quadrant,
> what happens? will alll segments of interest be actually
> filmed appropriately for analysis?
No, we will not be able to see all landmarks throughout the entire
activity (unless we use more than two cameras, which is possible, but in
practice presents problems of logistics, extra cost, and extra effort on the
part of the researchers). In the frames in which a landmark is hidden (for
instance, if we are digitizing the elbow joint, and it goes behind the trunk
for a few frames), we estimate its approximate position on the film as
closely as we can. But I think that such an estimation would be harder to
do when a SURFACE marker becomes hidden.
> would it not be better to get the same subject to perform
> under controlled conditions [eg in a laboratory], and thus
> use surface markers and the associated equipment?
The greatest problem would be that athletes often are not able to
repeat in non-competitive conditions the results that they achieve in
competitions. For instance, when Dick Fosbury won the Olympics in 1968 with
a 2.24 m high jump, the best that he had ever been able to clear in training
was 2.13 m.
> in situations like pathological gait analyses [eg for the
> elderly and infirm], i think that surface markers would be
> oe of the major ways to go. disagree? ;-)
I think you are probably right, in particular because in those
activities there will not be complex twisting movements to hide different
landmarks in different parts of the trial.
---
Jesus Dapena
Department of Kinesiology
Indiana University
Bloomington, IN 47405, USA
1-812-855-8407 (office phone)
dapena@valeri.hper.indiana.edu (email)