Removing noise in optical motion capture data

Question

Hello to you.
I am building my own optical motion capture system. In
relation to this I have coded a modified K-means clustering
algorithm in LINGO, which can be used for removing image noise in
the captured image data. Noise is in this case regarded as being
sparse groups of pixels (e.g they do not belong to a marker as
these are represented as compact pixel groups).
In order to evaluate the performance of my algorithm I need
to compare it with some other noise filtering solution. Does anyone
know if there is any suitable such LINGO code available? Possibly
Mean or Median filters…
I hope for positive replies:)

Answer 1

Jan Carlo wrote:
> Hello to you.
>
> I am currently trying to remove noise in image sequences
containing
> motion capture data. As a part of this project I am
clustering pixels
> into groups and want to remove pixel groups with less
than 4 members
> as these groups are regarded as noise.
>
> The problem is that when I have a nested list with e.g 3
sub
> lists/clusters (as indicated below) and I use a repeat
loop to remove
> the sub lists with less than 4 pixel coordinates, but my
code does
> not seem me to do this correctly (There are three sub
lists where 2
> of these should be removed, but only 1 of these 2 are
deleted while I
> am unable to remove the last one).
>
> Does anyone have any suggestions on how I can modify the
current code
> so that it does the job correctly?
>
The problem is that you're deleting things from the list
you're looking at
and things get moved towards the start, thus skipping an
iteration you
wanted to examine. If you start at the end, you won't get
that problem.
on test
clusters = [ [point(477, 453), point(477, 459), point(477,
460),
point(478, 454), point(478, 456)], [], [point(462,
405),point(462,
405),point(462, 405)]]
nMin = 4
nClusters=count(clusters)
repeat with i = nClusters down to 1
if clusters [ i ].count < nMin then
clusters.deleteAt(i)
end if
end repeat
put clusters
end test
Andrew