"I found a cute, funny, interesting, etc. movie"
Moderators: Stardust@home Team, DustMods
I have marked possible 'tracks' on two movies which turned to be calibration movies and has been judged wrong. It has resulted in worsening my specificity scores.
Based on what was taught in the tutorial, what I marked were possible tracks and needed to be flagged up. I am not aware if other participants have faced/facing the same problem.
This will inhibit me and maybe others from marking possible tracks on the real movies, which may influence the outcome of this project.
I would request the organisers to review the calibration movies to make sure any such errors are corrected.
I am sorry I was unable to go back to retrieve the ID of the movies which I had problems with.
Based on what was taught in the tutorial, what I marked were possible tracks and needed to be flagged up. I am not aware if other participants have faced/facing the same problem.
This will inhibit me and maybe others from marking possible tracks on the real movies, which may influence the outcome of this project.
I would request the organisers to review the calibration movies to make sure any such errors are corrected.
I am sorry I was unable to go back to retrieve the ID of the movies which I had problems with.
Did not work
Hi awmyth,
Answers to your question are at FAQ.
What I can tell you is carefully click right on the track, carefully click OK. Make it a long click.
Please see below.
http://stardustathome.ssl.berkeley.edu/ ... .php?t=473
fjgiie
Answers to your question are at FAQ.
What I can tell you is carefully click right on the track, carefully click OK. Make it a long click.
Please see below.
http://stardustathome.ssl.berkeley.edu/ ... .php?t=473
fjgiie
-
- Posts: 8
- Joined: Wed Aug 02, 2006 10:05 am
Marked a calibration movie
Hi Pollux.Castor,
Tell me about it ! I did the same thing.
The calibration movies should not have real tracks, only fake ones or no track, or out of focus. When we mark a track on a calibration movie that has "no track" then we are wrong. When we miss a track on a calibration movie that "has a fake track" then we are wrong. A calibration movie that is "out of focus" should be marked out of focus.
Since you and I marked a track on a calibration movie that "had no track" we were wrong.
We should learn from this that surface dots and such are not tracks, that the tracks are below the surface. Those glass beads also trip me up!
http://stardustathome.ssl.berkeley.edu/ ... e_number=4
Thanks,
fjgiie
Tell me about it ! I did the same thing.
The calibration movies should not have real tracks, only fake ones or no track, or out of focus. When we mark a track on a calibration movie that has "no track" then we are wrong. When we miss a track on a calibration movie that "has a fake track" then we are wrong. A calibration movie that is "out of focus" should be marked out of focus.
Since you and I marked a track on a calibration movie that "had no track" we were wrong.
We should learn from this that surface dots and such are not tracks, that the tracks are below the surface. Those glass beads also trip me up!
http://stardustathome.ssl.berkeley.edu/ ... e_number=4
Thanks,
fjgiie
Re: Marked a calibration movie
Hi fjgiie, hi Pollux.Castor,fjgiie wrote:Hi Pollux.Castor,
Tell me about it ! I did the same thing.
The calibration movies should not have real tracks, only fake ones or no track, or out of focus. When we mark a track on a calibration movie that has "no track" then we are wrong. When we miss a track on a calibration movie that "has a fake track" then we are wrong. A calibration movie that is "out of focus" should be marked out of focus.
Since you and I marked a track on a calibration movie that "had no track" we were wrong.
We should learn from this that surface dots and such are not tracks, that the tracks are below the surface. Those glass beads also trip me up!
http://stardustathome.ssl.berkeley.edu/ ... e_number=4
Thanks,
fjgiie
that negative callibration movie (2939826V1) hit me too. I agree to you Pollux.Castor: there was a very tiny (unknown) feature clearly subsurface.
[edit] oops: seeing the shiny dust in fjgiie´s sample on tutorial 4: must have been another one (former ID 9902) with a small dark feature"[/edit]
If you like, have a look at my "needle in a haystack" on page 4 of this thread. http://stardustathome.ssl.berkeley.edu/ ... =5056377V1
In fact there are three "needles":
1. a little "canyon" in the surface of the gel (I´m afraid much debris being located at the ground of such canyons is flagged as possible track)
2. a needle shaped artifact on the left side
3. a tiny "needle" on the bottom of the image just above the "o" of the scale (go to lowest focus). 9 out of 112 flagged that feature (I´d click again) thus having hardly the chance being reviewed by the scientific team.
Maybe some of us are "too" sensitive? Maybe the "sum" of us is not sensitive enough? Don´t know.
Cheers!
stardust1
Last edited by stardust1 on Sun Aug 13, 2006 1:31 pm, edited 2 times in total.
Wir leben in einer Zeit vollkommener Mittel und verworrener Ziele. (Albert Einstein)
Re: marked a calibration movie
stardust1,
I totally agree with several of your points.
That "needle" coming into focus at the bottom of the movie you posted (#3 in the list you made) is definitely interesting and I would mark it, too. I have marked a handful of such "tiny" interesting objects.
As I understand it, the system as it currently is biases toward missing tracks that size. If only 9 out of 112 people marked that as an object of interest, it seems like the Berkeley team may never even bother to check it out. Except of course, when it is rated as such by someone like stardust1 who has viewed 15,000+ videos and therefore gives weight to the judgment. But what about ones I marked, when I have only viewed about 600? They may never be looked at. If my opinion will be overlooked, I feel like, why do I bother?
Check out these:
near bottom right of this Movie 7322557V1: http://stardustathome.ssl.berkeley.edu/ ... =7322557V1
I see the surface about 12 bars from the top, while the speck comes into focus about 5 bars from the bottom. Only 6 out of 151 people have clicked it -- surely that pathetic rate will be attributed to noise and will be overlooked.
How about this one? Movie 5231889V1 - Near the middle of the bottom left quadrant.
http://stardustathome.ssl.berkeley.edu/ ... =5231889V1
12 out of 115 people have marked it. The surface is near the very top bar, while the speck comes into focus near the very bottom bar. What do you think??? Can there be dust contamination on the BOTTOM of the gel?
I admit that the video stardust1 posted looks even more exciting than these two, as it seems to have an impact-tunnel looking thing on the way down.
I think part of the problem is that the calibration vids are training us to catch objects that are 10-20 microns in diameter, according to the scale given (which is questionable as you have already pointed out). But the project's welcome page states http://stardustathome.ssl.berkeley.edu/ss_findingsd.php
"They are tiny—only about a micron (a millionth of a meter) in size!"
If they are expected to be a micron in size, why train us to observe 10-20 micron diameter tracks? It seems to me that people are basically being trained to overlook tracks of the size that is expected!
-Joot
I totally agree with several of your points.
That "needle" coming into focus at the bottom of the movie you posted (#3 in the list you made) is definitely interesting and I would mark it, too. I have marked a handful of such "tiny" interesting objects.
As I understand it, the system as it currently is biases toward missing tracks that size. If only 9 out of 112 people marked that as an object of interest, it seems like the Berkeley team may never even bother to check it out. Except of course, when it is rated as such by someone like stardust1 who has viewed 15,000+ videos and therefore gives weight to the judgment. But what about ones I marked, when I have only viewed about 600? They may never be looked at. If my opinion will be overlooked, I feel like, why do I bother?
Check out these:
near bottom right of this Movie 7322557V1: http://stardustathome.ssl.berkeley.edu/ ... =7322557V1
I see the surface about 12 bars from the top, while the speck comes into focus about 5 bars from the bottom. Only 6 out of 151 people have clicked it -- surely that pathetic rate will be attributed to noise and will be overlooked.
How about this one? Movie 5231889V1 - Near the middle of the bottom left quadrant.
http://stardustathome.ssl.berkeley.edu/ ... =5231889V1
12 out of 115 people have marked it. The surface is near the very top bar, while the speck comes into focus near the very bottom bar. What do you think??? Can there be dust contamination on the BOTTOM of the gel?
I admit that the video stardust1 posted looks even more exciting than these two, as it seems to have an impact-tunnel looking thing on the way down.
I think part of the problem is that the calibration vids are training us to catch objects that are 10-20 microns in diameter, according to the scale given (which is questionable as you have already pointed out). But the project's welcome page states http://stardustathome.ssl.berkeley.edu/ss_findingsd.php
"They are tiny—only about a micron (a millionth of a meter) in size!"
If they are expected to be a micron in size, why train us to observe 10-20 micron diameter tracks? It seems to me that people are basically being trained to overlook tracks of the size that is expected!
-Joot
Re: marked a calibration movie
No, I don´t think so. It was told, the ranking of the movies would be weighted appropriate to the stats of those who found it. I don´t think they did at the 1st review. Look at this: http://stardustathome.ssl.berkeley.edu/ ... .php?t=797joot wrote:... As I understand it, the system as it currently is biases toward missing tracks that size. If only 9 out of 112 people marked that as an object of interest, it seems like the Berkeley team may never even bother to check it out. Except of course, when it is rated as such by someone like stardust1 who has viewed 15,000+ videos and therefore gives weight to the judgment. But what about ones I marked, when I have only viewed about 600? They may never be looked at. If my opinion will be overlooked, I feel like, why do I bother?
And if so, hopefully they would regard accuracy first.
And: Any of your finds will be looked by other volunteers again and again. I guess any movie will at least be seen by about the same number of volunteers ind the "top 100" list (maybe 1-3).
Hopefully the scientists will not have to look at all those kind of finds. Otherwise the whole project would make no sense. Maybe they will have a closer look, if they don´t find enough particals in the 1st review.
So don´t be discouraged.
About the movies you posted (including my possible find):
I would not say, those "tiny unknown features focusing below the surface" are promising in any kind. How could I? I´m not the expert to do. It´s just: at the moment there is no evidence given to me to flag them not. Seeing your samples, there are some comparable in the list of "my events". Those reach confirmation rates of about 20-40%. Maybe in some weeks scientists will realize and say: "hey folks, don´t flag such things furthermore. There is nothing behind."
Don´t forget there are supposed to be only some dozens of particles (which could be a strong indication not to flag such finds. But who knows? And don´t forget all that particals have to be retrieved from the aerogel at the end. So it´s obvious to search for the "bigger" ones with first priority.
We don´t look at particles, we look at their tracks which are much bigger. (please have a look at the tutorial)... I think part of the problem is that the calibration vids are training us to catch objects that are 10-20 microns in diameter, according to the scale given (which is questionable as you have already pointed out). But the project's welcome page states http://stardustathome.ssl.berkeley.edu/ss_findingsd.php
"They are tiny—only about a micron (a millionth of a meter) in size!"
regards
Last edited by stardust1 on Mon Aug 14, 2006 2:18 pm, edited 1 time in total.
Wir leben in einer Zeit vollkommener Mittel und verworrener Ziele. (Albert Einstein)
I call this one 'Cleopatra with Cigarette'
(9 people clicked this one so far)
http://stardustathome.ssl.berkeley.edu/ ... =5670244V1
(9 people clicked this one so far)
http://stardustathome.ssl.berkeley.edu/ ... =5670244V1
Hey alexz,
nice find! Thank´s a lot!
Must be because I´m blind on the obvious and/or unimaginative:
1. to me that´s Nofretete
wife of Echnaton (the lucky)
the most beautiful woman ever been portrayed in the history of mankind
without cigarette
2. where did those 9 click? confused
Cheers!
nice find! Thank´s a lot!
Must be because I´m blind on the obvious and/or unimaginative:
1. to me that´s Nofretete
wife of Echnaton (the lucky)
the most beautiful woman ever been portrayed in the history of mankind
without cigarette
2. where did those 9 click? confused
Cheers!
Wir leben in einer Zeit vollkommener Mittel und verworrener Ziele. (Albert Einstein)
-
- Posts: 8
- Joined: Wed Aug 02, 2006 10:05 am