Stardust@home Options

[Rakhir]

I don't found any post about Stardust@Home so I created a new thread.

Stardust@Home is a distributed search by volunteers for interstellar dust in the Stardust interstellar dust collector.

Volunteers have to pre-register for an expected start in spring of 2006. You will also have to pass a test to be qualified for the search.
Indeed, unlike the distributed computer projects running in background tasks, this project is using your eyes to scan "focus movies" thanks to a Virtual Microscope.

According to estimations, there should be about 45 interstellar dust impacts in the collector.

Besides the satisfaction of contributing actively to this sample return project, your name will appear as a co-author on the paper announcing the discovery of the particle.

http://stardustathome.ssl.berkeley.edu/

Rakhir

[deglr6328]

Looks fun. What do they mean though when they say "Not even one contemporary interstellar dust grain has ever been studied in the laboratory"? Were old grains found somewhere?

[hendric]

http://www.psrd.hawaii.edu/April00/analyzingSmall.html

has some good information. I think what they mean by contemporary is "recent". Grains studied so far have been from meteorites etc, dating back to the birth of the solar system.

[Ames]

Pre-registered and ready for action.

I think this is an excellent opportunity for UMSF members to actively participate in a research project. Ok it is a monotonous search for a tiny mote of dust, but much science is just like that, and just think how you will feel if you are one of those who found one of the 45.

Nick

[akuo]

Is there some sort of explanation from the Stardust team why this detection cannot be done by a computer program scanning the images? Are the marks left by interstellar dust so unpredictable that human work is needed? How does the person know what to look for then?

I'd imagine that the trail would just be longer because of higher velocity of interstellar particles, so it seems strange that this couldn't be detected by a machine.

Anyhow, I already signed for the job anyway :-)

[djellison]

It's probably cheaper to get people to do it than write some software to do it Especialyl when the people are a) free and plentiful

There's going to be some 'training' images before you're allowed to do real ones as I understand it, to make sure you know what you're doing

It reminds me of the Mars crater counting on 'clickworkers'

Doug

[paxdan]

stardust@home

please be mac compatible please be mac compatible please be mac compatible please be mac compatible please be mac compatible please be mac compatible

that is all.

[elakdawalla]

It's probably cheaper to get people to do it than write some software to do it   Especialyl when the people are a) free and plentiful

There's going to be some 'training' images before you're allowed to do real ones as I understand it, to make sure you know what you're doing

It reminds me of the Mars crater counting on 'clickworkers'

Yes, it's something like clickworkers. I beta-tested their training materials and it is tedious but I could get hooked anyway and I think it is a Web browser-based Java application, so it should be platform-independent, provided you've got Java working properly on your machine. We're collaborating on the project, mostly to try to help them get the word out to people. They'll have an awful lot of images to look at!

--Emily

[djellison]

I did WAYYYyy too much crater clicking...it was strangely addictive.

I'm guessing the 'movies' refered to in that TPS article are a pull of focus thru the gel?
Doug

[odave]

...I think it is a Web browser-based Java application

Do you know if your connection speed will have an impact on performance?

Since I really doubt my company would pay for my time & bandwidth doing stardust@home (silly, I know, but they do keep going on about profit and such), I'll be doing it from my dialup access at home.

This could be the straw that breaks the camel's back for my getting broadband at home, though...just need a way to get it through the Household Purchasing Department

[elakdawalla]

I'm guessing the 'movies' refered to in that TPS article are a pull of focus thru the gel?

Yup. You catch the dust particle paths because they are beneath the surface, not on the surface, so they pop into focus when everything on the surface is blurry. I think that the "movies" are really just stacks of a couple dozen individual JPG images, which you can click forward and backward through, and they're not very high resolution, I think maybe only 400 or 500 pixels wide. So it should be possible (though not speedy) to work on this through a dialup.

--Emily

[The Messenger]

Is there some sort of explanation from the Stardust team why this detection cannot be done by a computer program scanning the images? Are the marks left by interstellar dust so unpredictable that human work is needed? How does the person know what to look for then?

I'd imagine that the trail would just be longer because of higher velocity of interstellar particles, so it seems strange that this couldn't be detected by a machine.

Anyhow, I already signed for the job anyway :-)

It is extremely difficult to train a computer to recognize and discriminate fringe events - 1 to eight pixels. While the human eye can quickly discern a pattern of shadows (due to variation in light, or slight density changes in the medium), a computer will often identify these as false positives.

Another problem is patterns in a neutral medium that are periodic but slightly asymetric. These are flagged as contaminants by virtually any pattern recognition software. I can eyeball a slice of Swiss cheese, and easily tell if a hole formed naturally, or if it was left by a core sampler. A computer would have fits trying to tell the difference.

There is a good probability that all of the images will be prescreen by computer, and only images with possible inclusions will be distributed to image evaluators.

(FWIW, I have written neuro-network-like routines that are used to evaluate rocket motor propellant grains for many programs...fun stuff, and if the medium is right, I might be able to turn them loose on Stardust.)

[ElkGroveDan]

stardust@home

please be mac compatible please be mac compatible please be mac compatible please be mac compatible please be mac compatible please be mac compatible

that is all.

They have audio tape explaining how to convert the program for your mac. It's an 8-track.

[dvandorn]

I registered with Stardust@home as soon as its existence was revealed on the forum, here. I've yet to hear anything at all back from them.

Has anyone heard from them? Or did they only take the first couple of thousand applications, and I'm SOL?

-the other Doug

[Rakhir]

I registered with Stardust@home as soon as its existence was revealed on the forum, here.  I've yet to hear anything at all back from them.

Has anyone heard from them?  Or did they only take the first couple of thousand applications, and I'm SOL?

-the other Doug

I also registred the same day I created this thread but received no news from them.

Rakhir

[djellison]

No news yet - I doubt they'll send out emails for the 'test' until they have proper imagery ready to go - so end of this month, beginning of next I'd guess

Doug

[tty]

Is there some sort of explanation from the Stardust team why this detection cannot be done by a computer program scanning the images? Are the marks left by interstellar dust so unpredictable that human work is needed? How does the person know what to look for then?

Because people have an image processing and pattern recognition software package that has had about 1,000,000,000 years of development work invested in it.

tty

[Guest_PhilCo126_*]

Well, with Stardust (finally) at home... I remember reading something that the rest of the capsule might be displayed at the Natioanl Air & Space Museum in Washington D.C.

[ljk4-1]

Well, with Stardust (finally) at home... I remember reading something that the rest of the capsule might be displayed at the Natioanl Air & Space Museum in Washington D.C. 

This got me thinking: What will they eventually do with the Genesis return capsule?

[ElkGroveDan]

Because people have an image processing and pattern recognition software package that has had about 1,000,000,000 years of development work invested  in it. 

With open source algorithms!

[Guest_BruceMoomaw_*]

This got me thinking:  What will they eventually do with the Genesis return capsule?

Sweep its pieces quietly under the rug.

[djellison]

TPS members will be pleased to see that you can now do the 'beta' testing phase - interface is actually very good, I'm enjoying it.

Your score: 10 out of 10

Passing score: 8 out of 10

I'm a pro, me.

Doug

[Guest_BruceMoomaw_*]

An LPSC abstract on the results from the first test of Stardust@Home's likely reliability shows promising results: http://www.lpi.usra.edu/meetings/lpsc2006/pdf/2225.pdf . (There's also a description of it from the previous year's LPSC: http://www.lpi.usra.edu/meetings/lpsc2005/pdf/1908.pdf .)

[Borek]

With open source algorithms!

Show me the source, then.
I would say that the sources are very proprietary and hard to reverse-engineer.

Borek

[Bob Shaw]

This got me thinking: What will they eventually do with the Genesis return capsule?

They'll display it in 23 museums across the US of A!

Bob Shaw

[Guest_BruceMoomaw_*]

In a bag...

[djellison]

To be fair, the samples may be scrap, but the aeroshell is recognisable - but I don't think they'd want to display it by any stretch of the imagination. It would be like exhibiting the dent in your bumper before getting an insurance quote for your car.

Doug

[helvick]

I would say that the sources are very proprietary and hard to reverse-engineer.

But really easy to copy, well replicate at any rate with some modifications

[MahFL]

I passed the Stardust@Home training, 10/10

[Bob Shaw]

To be fair, the samples may be scrap, but the aeroshell is recognisable - but I don't think they'd want to display it by any stretch of the imagination. It would be like exhibiting the dent in your bumper before getting an insurance quote for your car.

Doug

Doug:

Actually, all the more reason to display it - the debris should be preserved in the (several) places where the errors were made which resulted in the parachute failure, along with a full presentation about the causes, the enquiries, and the lessons learned in terms of physical and management procedures!

But they'll probably quietly lose it.

Bob Shaw

[tty]

Show me the source, then.
I would say that the sources are very proprietary and hard to reverse-engineer.

Borek

The source code is open, but the documentation is lousy.

tty

[dvandorn]

I finally received an e-mail saying that Stardust@home is launching. The e-mail gave a link to the Berkeley people who operate the various distributed processing projects out there.

Unfortunately, the site is non-functional -- the banner links to actually get started working on the project aren't working (aren't even links, are just plain text), and the message says that the website might be either slow or unresponsive because of the high volume of traffic.

I'm hoping the links start working in the next day or two -- I'm anxious to get started!

-the other Doug

[Tman]

Got that e-mail last night too - And after celebrating finally our national holiday (pang, boom, whoosh... ) I was able to pass the test about the ten VM-movies and then to create an account (and to print my Planetary Society Certificate ).

Actually I had to start with the test twice because the website also went down for some time. Then later I was able to log in and start with some real VM-movies (1...25) until I got a lock-out again.

As far as I can see I only found (or got the) tracks in the sporadical appearing calibrated movies and ranked (circa) on the 350 of 1880 place...

[Tman]

Latest News: Aug. 2, 2006 - 7:26am

We have shut down the training, testing and VM section of the Stardust@home website because of a problem in which random images of unknown origin appear in the focus movies...

http://stardustathome.ssl.berkeley.edu/for...topic.php?t=172

I saw a wedding group picture too.

[dilo]

An hacker attack ongoing?!?

[RobR]

An hacker attack ongoing?!?

I think they were just overwelmed. They are up again now and performing very well.

Statistics for robreeve

Your Overall Score: 96
Total Movies Viewed: 410
Your Rank: 1062 out of 5411
Specificity: 100%
Sensitivity: 89%

[I have to work on that sensitivity] <grin>

Stardust found:

Particle number Total viewings Number of Agreements Official status Stardust@Home Team OfficialComment
Movie 39071 14 3 Indeterminate
Movie 20962 15 3 Indeterminate
Movie 41699 17 4 Indeterminate
Movie 43214 16 4 Indeterminate
Movie 22712 17 2 Indeterminate

[Tman]

Hi Rob, welcome to UMSF!

Statistic for "greuti":

Calibration Movies Answered Correctly 471
Calibration Movies Answered Incorrectly 3
Your Overall Score: 468
Total Real Movies Viewed: 1230
Your Rank: 45 out of 5641
Specificity: 100%
Sensitivity: 99%

It seems to be a form of race There're some searchers with a lower score in front of me - I guess they've still no failings

Slowly I need a real feeling of success by finding a real (new) track instead of these calibration movies... I hope that's already a real search and not merely another screening of serious searchers...

[RobR]

1230 Movies! You should have some events there, click the "My events" tab. If I found 5 possibles in 300, you shluld have more after 1230....

[Tman]

Yes I clicked two at the beginning (status "Indeterminate" too), but meantime I think that aren't real tracks or particles (later I've still found more of them without clicking on).

What I find is that a lot of movies are not in good focus (with little scope below the surface). In such movies you wouldn't see the track as defined ring but at most as faint dark spot or a form of shadow.

[dilo]

What I find is that a lot of movies are not in good focus (with little scope below the surface). In such movies you wouldn't see the track as defined ring but at most as faint dark spot or a form of shadow.

Hi Tman! Usually I categorize "bad focus" these images (perhaps you're smarter than me )... sometimes entire movie is out of focus!
Anyway, congrats for the 1230 movies!!! Where do you find all such time? Or, again, you are faster/smarter than me???

Edit: I also found two very small tracks (status indeterminate too, most people didn't see anything)

Statistics for dilo
Your Overall Score: 25
Total Movies Viewed: 114
Your Rank: 1955 out of 6110
Specificity: 94%
Sensitivity: 85%

[centsworth_II]

The stardust info I read said that there were expected to be "several dozen" particles trapped in the collector. If there are about 100 particles to be viewed in the eventual 700,000 focus movies, that would be about one per 7,000 movies. It would be a shame to view thousands of movies and then miss the one particle that was present in them.

I recall there were to be test movies presented every so often similar to the instructional movies containing examples of particles. I wonder how often. I would think there are more test particle images than real particles scattered amongst the movies. I wonder if they will tell you if you have identified a test particle image.

[dilo]

I wonder if they will tell you if you have identified a test particle image.

No way, they report as "my event" only real particles observed in Stardust samples.
FYI, the incidence of "calibration movies" is 25-30% of the total (based on my experience).

[Tman]

If you click on a movie (test or real) including a track and your score dosen't increase then you probably found a new track. Btw. the adress in your browser shows (so far) whether it's a test movie or a real movie

@dilo, fortunately, I work currently temporary

[RobR]

By the way, they advertise on the site that you will get 20% test movies.

Statistics for robreeve

Your Overall Score: 117
Total Movies Viewed: 524
Your Rank: 1062 out of 6476
Specificity: 100%
Sensitivity: 88%


Movie 39071 21 3 Indeterminate
Movie 20962 21 4 Indeterminate
Movie 41699 30 6 Indeterminate
Movie 43214 26 8 Indeterminate
Movie 22712 24 3 Indeterminate
Movie 41954 17 4 Indeterminate
Movie 41322 32 6 Indeterminate

[odave]

Heh, here's my stats - just signed up:

Your Rank: 65535 out of 6543
Specificity: unknown
Sensitivity: unknown

How's that for an ego hit when your rank is so far below the number of participants!

This is fun, and not too bad on my dialup connection. The movies are taking ~3 minutes to come down, so it's not too bad. It'll take a while to get up over 1000, though.

[RobR]

I took you comment obout the URL telling test or real - and almosst immediately, got one that said real, found an obvious hit, and then my test hit went up by one. Looks like maybe they're onto ppl.

[RobR]

I took you comment obout the URL telling test or real - and almosst immediately, got one that said real, found an obvious hit, and then my test hit went up by one. Looks like maybe they're onto ppl.

OIC! The URL shows the on-deck movie - the next one.

[dilo]

Did somone noticed these 3 small particles (light blue arrows) which appear always focused in all movies.

I think could be something on the microscope optics... anyway dangerous for false identifications, at least for a beginner!

[dvandorn]

Finally got in and going, myself. Had an interesting one -- didn't take a screen cap or anything, since I just go along, give it my best shot, and go on to the next...

Anyway, there was one image that appeared to be pretty flat (no big slope on the subject), but two very, very small dark spots, maybe 1.5 microns or so based on the scale bar, came into focus at the deepest level. No track visible in any way, just a dark spot deep inside the aerogel. I marked the bigger of the two spots as a track, but that was a really soft identification.

I've also run across several images that do come into focus, but the perfect surface focus point is at the deepest focus setting on the adjustment bar. I reported those as "bad focus," since I figure the point of the exercise is to zoom your focus down into the aerogel. Any movie that doesn't let you get below the surface is, I figure, guilt of a focus problem.

Are y'all seeing the same kind of things?

Seems like almost all of the tracks I've seen are the ones in the test images... though my scoresheet says I've identified something like six tracks, and while a couple of them only have corroboration by one or two other people (out of 20 to 30 views), some of them have been pegged by a dozen or more people.

It's fun, and I feel like I'm actually doing something useful!

-the other Doug

[odave]

Did somone noticed these 3 small particles (light blue arrows) which appear always focused in all movies.
I think could be something on the microscope optics

I've seen them too, and I think you're right. One of the tutorial movies points out something similar as being dust on the camera.

[Tman]

Doug, if you like to show what you've marked/stored in your "My Events" then you can copy the URL like this:
http://stardustathome.ssl.berkeley.edu/mye...?movie_id=43887
I think it shows something similar to that you've described...

Above we've already touched on the focus problem. I report those as "bad focus" too - there're a real lot of them


@dilo, I noticed they too. It's first something confusing but they do not change their form.

[jaredGalen]

Seeing as we are posting stats....
My rank hasn't changed in ages. Guess I'm an average user so I stay pretty static.

Your Overall Score: 82
Total Movies Viewed: 283
Your Rank: 2697 out of 6787
Specificity: 100%
Sensitivity: 100%

Movie 45071 41 3 Indeterminate
Movie 32403 16 1 Indeterminate
Movie 20236 29 3 Indeterminate
Movie 34526 29 2 Indeterminate
Movie 54510 46 1 Indeterminate
Movie 44387 32 7 Indeterminate
Movie 20324 35 6 Indeterminate

[deglr6328]

I'm having a terrible time with bad focus images. I've done maybe 100 "movies" and at least a third have the surface in focus WAY far away from the center of the slider bar. Some movies are so devoid of any features at all that I can't tell if I'm actually looking at different images or if the javascript has frozen up and I'm stuck on the same one. If we're looking for ~10 tracks in ~1 MILLION images I have little hope the method will deliver real results with this kind of performance. The images should contain a reticle fiducial like this which rests on the srface so the viewer can see immediately where the surface is.

It looks to me like all they've done is get a block of aerogel under the microscope, focus to the surface and tell the autoscanner/imager "ok here's the surface, scan the whole block assuming this level at xx microns will be where the surface is" regarless of surface level deviation, when what they should be doing is surface level calibrating every movie made by using a little piece of software to see where the highest contrast is (like a digital camera does) and or using a small/thin surface fiducial at the image edge. Also Dilo is right about those dots that appear on every image, it makes me think they didn't even flatfield calibrate the CCD on the imager before they started!! not good. On the other hand the user interface is pretty nice and natural to use.

[RobR]

I'm having a terrible time with bad focus images. I've done maybye 100 "movies" and at least a third have the surface in focus WAY far away from the center of the slider bar. Some movies are so devoid of any features at all that I can't tell if I'm actually looking at different images or if the javascript has frozen up and I'm stuck on the same one. If we're looking for ~10 tracks in ~1 MILLION images I have little hope the method will deliver real results with this kind of performance. The images should contain a reticle fiducial like this which rests on the srface so the viewer can see immediately where the surface is.

It looks to me like all they've done is get a block of aerogel under the microscope, focus to the surface and tell the autoscanner/imager "ok here's the surface, scan the whole block assuming this level at xx microns will be where the surface is" regarless of surface level deviation, when what they should be doing is surface level calibrating every movie made by using a little piece of software to see where the highest contrast is (like a digital camera does) and or using a small/thin surface fiducial at the image edge. On the other hand the user interface is pretty nice and natural to use.

Bah, Humbug you say.

I've been 550 'real' slides now and I don't see that big a problem with the poorly focused examples. Just click bad focus and go on. the odds are low but... you might have one named after you!

my stats:

Your Overall Score: 186
Total Movies Viewed: 761
Your Rank: 885 out of 7330
Specificity: 100%
Sensitivity: 89%

Movie 20962 31 5 Indeterminate
Movie 41699 41 9 Indeterminate
Movie 43214 35 11 Indeterminate
Movie 22712 37 3 Indeterminate
Movie 41954 37 7 Indeterminate
Movie 41322 47 7 Indeterminate
Movie 43323 49 5 Indeterminate
Movie 41590 46 10 Indeterminate
Movie 41491 48 1 Indeterminate
Movie 44039 28 4 Indeterminate

[Tman]

Something about the small dark spots below the surface:

http://stardustathome.ssl.berkeley.edu/for...topic.php?t=476


@About the huge amount of "bad focus" movies, I bet they are working already on a better solution to approach and scan again those positions - a lot additional work though.


@Rob, its not needed to quote the lot when you reply right after it.

[Tman]

Warning!

http://stardustathome.ssl.berkeley.edu/for...asc&start=0

It cost me at least 5 points with the same growing of my incorrect answers and a clean 100% 100%!
And it's not yet well fixed so far!

It's not important for the search - but angry for my glory ;-)

[RobR]

I've had quite a few failed calibration movies (15) which I've looked at exhaustively and have not been able to 1. find my mistake or 2. I found the freaking track and got it wrong anyway. I stopped paying attention to that score, too fustrating to track; If they can't get it right I'm not going to worry anymore.

However, I can't help but wonder about the overall impact on the project, i.e.if ppl don't pay attention to their score, the testing program becomes ineffective, and quality control metrics go in the tank.

Your Overall Score: 210
Total Movies Viewed: 870
Your Rank: 925 out of 8146
Specificity: 100%
Sensitivity: 88%

Movie 39071 37 4 Indeterminate
Movie 20962 36 5 Indeterminate
Movie 41699 52 11 Indeterminate
Movie 43214 47 16 Indeterminate
Movie 22712 55 3 Indeterminate
Movie 41954 52 12 Indeterminate
Movie 41322 62 10 Indeterminate
Movie 43323 66 7 Indeterminate
Movie 41590 53 11 Indeterminate
Movie 41491 60 2 Indeterminate
Movie 44039 36 4 Indeterminate

[tty]

Stardust-at-home's early travails have made it into Nature:

http://www.nature.com/news/2006/060731/full/060731-10.html


tty

[dilo]

Very strange structure on movie 845194V1:

The particle on the top show an intriguing radial edge, like an impact crater; however, structure seems located on the surface (I report 3 movies images focalized respectively above, within and below surface from left to right). I didn't cataloghed as a track, but I still confued.

[Tman]

Yea, there're some odd structures appearing at times. I guess it could be such a thing that according http://stardustathome.ssl.berkeley.edu/ss_...hedule_number=5 are not well-understood so far.

@about the calibration movies which had incorrect coordinates in them for the locations of the tracks (and gave you an incorrect answer back). Now it seems definitely fixed since monday. Also all movies got new IDs since then. So behind your first 10'000 movies, there should be no more impact to your score list (if you've scanned correctly through all of them)

You may be a starduster if...
- you've started giving names to the dust in your house
- instead of looking for your glasses in the morning, you try to find the focus bar
- you're getting worried when you don't see Larry, Curly and Moe on your car windshield
- there's a gutter in your mouse mat from rolling over the same two inches over and over again
- when you're channel flipping and an old episode of The Simpsons comes on, you think "obviously a calibration movie"

I guess "Larry, Curly and Moe" mean these 3 small (camera dust) particles which appear always focused in all movies.

[Tman]

Another strange (but nice) thing:

[Guest_AlexBlackwell_*]

Andrew Westphal is "glogging" for Emily this week. Check out his first entry.

[CosmicRocker]

sorry about the flood...

Yeah, I just caught up with her blog today and noticed his informative entry. I wish he would have given a little more practical information for us searchers. After viewing a bunch of movies I am still a bit confused about some things. Still, his glog entry is the best so far. The first guy was, well, how should I put it? I'll bite my tongue. The second was better, but still not on the mark of what I had come to expect from Emily. If Andrew fills us in with some better information than the stardust@home site and forum has so far provided, he'll be a big hit with me. I think we all are waiting for one of the gloggers yet to come. I haven't found one, but is there a thread on that topic yet? I was going to start one several times, but chickened out.

It's unfortunate that we haven't been given more guidance as this stardust thing has developed. Almost all of the anomalies I've seen have been inclusions that come to focus below the surface. One thing that is a real mystery to me is why all of the calibration tracks appear in focus only below the aerogel surface. I can understand why they remain in focus below the surface, but shouldn't they start _at_ the surface?

Anyway, so far I am tagging anything that comes to a focus below an identifiable surface. There are a lot of others agreeing with my clicks, but I am sure they are not tracks. I haven't seen any real movies that resemble the calibration tracks. My search method is this... Evaluate the general focus of the movie, if possible identify the surface, determine the orientation of the surface (is it tilted in some direction, or does it have a more complex shape?), look for stuff that comes to a focus beneath any indentified surface.

I've been getting a bit burned out on the movie viewing lately. If we got some feedback from the officials regarding some real tracks found, I could find some new energy. It is tough to view a lot of movies on dialup at home, and using my broadband connection at work was seriously affecting my productivity there. I had to consider who was writing my paychecks. I was hanging with the top 1000 until I had to make that hard decision. As for the top 100, I'll bite my tongue once more.

After 1400+ total movies viewed, my rank is now lagging. It seems a shame that ranking seems to depend only on total movies viewed. I would have thought accuracy should be part of the equation. I think they are giving users the wrong feedback. I really could care less about the ranking, but I really want to identify a track. Perhaps they are really serious when they say "expect the unexpected." Although they theoretically expected only about 45 tracks in the whole array, I can only guess that they have not positively identified one yet.

It hasn't been terribly active lately, but earlier there were more helpful comments posted in the freenode IRC channel #stardust@home...

My current stats, which are not changing much lately...

Calibration Movies Answered Correctly 339
Calibration Movies Answered Incorrectly 3
Your Overall Score: 336
Total Real Movies Viewed: 1082
Your Rank: 1351 out of 11176
Specificity: 100%
Sensitivity: 98%

[odave]

I realize this post is probably more appropriate at the stardust@home forums, but they're still a bit of a mess, IMHO. So for my fellow UMSF dusters...

One thing that is a real mystery to me is why all of the calibration tracks appear in focus only below the aerogel surface. I can understand why they remain in focus below the surface, but shouldn't they start _at_ the surface?

I've wondered about this too. You may have noticed (on some of them) that before you get to the "soda straw" cross-sections, the aerogel tends to have a hazy/cloudy appearance just over it. I think this is the wide part of the carrot-shaped track, with the narrower part of the carrot being the in-focus soda straw, below the surface.

The tutorial shows a nice, circular edge for the beginning of the track, but based on this image, I think the top is much more messy. Here's another image.

It does seem odd to me that we don't see more disturbance at the surface over a track - my other theory was that the aerogel movies are upside down from the POV of the particle, but the documentation seems to disprove that.

Anyway, so far I am tagging anything that comes to a focus below an identifiable surface.

Me too. Given that these particles are thought to be tiny, I think that anything that looks different should be marked for closer inspection. I figure that's why we're participating - the team can then use the "number of agreements" stat to filter and prioritize reviews.

It seems a shame that ranking seems to depend only on total movies viewed. I would have thought accuracy should be part of the equation.

Accuracy is in the ranking equation, but it is definitely influenced by the number of movies viewed. From the definitions page:

Score

Your score is the total number of calibration movies you have identified correctly minus the number of calibration movies you have answered incorrectly. Most movies are not calibration movies, and will not affect your score. So don't be surprised if your score stays at 0 for a while.

So from the current rankings page, rhar006 is #1 with a score of 9963. This means that they have answered 9963 calibration movies correctly. To get that high, rhar006 needs to be both very accurate and extremely dedicated (read:has no life ), as I am getting about 30% calibration movies.

I'm with you though, I worry more about my accuracy percentages than where I'm at in the ranks. I'd rather look at fewer movies and have the team see that I know what I'm looking for.

FWIW:

Statistics for odave

Calibration Movies Answered Correctly 303
Calibration Movies Answered Incorrectly 2
Your Overall Score: 301
Total Movies Viewed: 1000
Your Rank: 1534 out of 11253
Specificity: 100%
Sensitivity: 99%

[odave]

FYI - there's a new update over at the stardust@home forums about the "interesting candidates" discovered so far. The two highest scored movies they list have features in them that look just like tracks on the calibration movies - but it's way to early to tell for sure. But still, very cool! Unfortunately most of my events are of the "black speck" variety, no soda-straws here yet

[Tman]

Yea, but in one of the two highest scored is the focus sure above the surface. There're many of such surface "circles"/structures among the movies.

The third is really interesting. If it's a track I wonder why it could broaden again.
(Edited: oops I guess it dosen't broaden, it's a focus issue)

Another interesting update: http://stardustathome.ssl.berkeley.edu/for...topic.php?t=874
As expected but even so astoundingly few such "cheaters" among the top 200.

Now after a considerable quantity of searching, one observation is really a bit questionable to me: I'm sure I got several Real Movies at least twice and had in addition some déjà vu feelings. So I wonder why happens that at all. Shouldn't be the system so designed that each movie is checked off to you when you've made a decision? Mainly in view of the huge amount of movies which are still to scan - and personally I don't feel like searching through twice

[MichaelT]

Some of my candidate movies have been checked and were found to be "Not Extraterrestrial" "inclusion". They are these dark spots that many others have spotted as well. So far I have only found such dark spots and I'm afraid that none of my candidates are of extraterrestrial origin.

Movie 3316648V1 158 40 Not Extraterrestrial inclusion
Movie 3799221V1 158 38 Not Extraterrestrial inclusions
Movie 4369380V1 169 37 Not Extraterrestrial inclusions

Michael

[Tman]

Hi Michael, nice to see you here!
Yea, these dark "particles/inclusions" were a "funny" Stardust@home "story"
Meantime I do no more click they too.

As far as I know, there're some possible tracks (even with the particle) but so far all like this example: http://stardustathome.ssl.berkeley.edu/ss_...edule_number=10
To find in the "I think I've found a track, what do you think?" thread:
http://stardustathome.ssl.berkeley.edu/for...topic.php?t=647


Btw. I found the following movie of my teeny "interplanetary sun-glider"




Regarding the real movies which appear twice or repeated. It seems all real movies get added to the same pool and then random selected to the VM. If so I hope that they retire them when they get past a certain number of viewings (and decisions) by different searchers.

[Tman]

Such a possible track with the particle (that I found few days ago too in ~36 place): http://stardustathome.ssl.berkeley.edu/mye...php?id=433711V1

I just saw in "My Events" that it passed cut 1 and got the official comment "another IDP" (does anybody know what IDP means?).

[CosmicRocker]

odave, thanks for mentioning that update about the "interesting candidates." It's good to finally get some kind of feedback. I guess I'll stop selecting the inclusions also, but so far none of my movies have been reviewed yet.

[ugordan]

(does anybody know what IDP means?).

Interstellar Dust Particle?

[Tman]

Interstellar Dust Particle?

(most) likely

[odave]

From albutterworth's update, it's Interplanetary Dust Particle. That is,

It is an impact track, almost certainly not interstellar. It's trajectory will tell us whether it is likely to be secondary ejecta from the Stardust spacecraft solar panels, or an Interplanetary Dust Particle (IDP).

The stardust@home definitions page is the official glossary, but IDP hasn't made it on there yet - they are updating it as they go.

I've finally had one of my movies get marked as reviewed - "passed cut 1: weird WCO" (Worth Checking Out), so they are making progress...

[Tman]

Oh yea, thanks! I've read that article, but it seems not complete

Btw, this movie appears currently in my Events as "159 / 41 / Passed cut 1 / another IDP".
Would say rather few of 159 that clicked on it.

I've finally had one of my movies get marked as reviewed - "passed cut 1: weird WCO" (Worth Checking Out), so they are making progress...

Very weird if it's below the surface.

[Tman]

There's a nice example of another type of impact track in the aerogel besides the straight tracks:



The focus is circa left of center of the track.
Yesterday I found (in fifth place - of already 150 viewings!?) the right part of it and posted it right in the stardust forum. Anna (Stardust@home team) was so kind to point out that they've already found the left part of it (should I've known that - I should ).

[odave]

One thing that is a real mystery to me is why all of the calibration tracks appear in focus only below the aerogel surface. I can understand why they remain in focus below the surface, but shouldn't they start _at_ the surface?

In case anyone's interested, there's a new thread on the Stardust@home forums that discusses this.

[Tman]

It seems there're no longer Starduster here

Odave (or someone else) what do you think of my interpretation I've posted there? (Also because I don't know how coherent my "English syntax" is)

Have a look at that image too that shows two tracks from particles which entered in the areogel by supersonic speed (which is also expected to interstellar particles): http://www.planetary.org/image/PIA02188.jpg

My interpretation of that we're looking at the tracks in the areogel is, the black distinct circle appears just around the widest points of the track because it's that point where the black border appears most distinct when we look straightly at it from above. In that view the "blurry section" is the distinct rounded border just beneath the surface. That explain also the unseeable bullet hole direct at the surface because there the track border is thinnest to our view.

Additional it could be the interstellar particle was that tiny, that it caused a bullet hole at the direct surface which is either way too small for watching it in that magnification.

[odave]

I agree with your interpretation of the tracks, TMan - and your syntax is just fine.

I'm still dusting, though only during my lunch break at work. I have dialup at home, and while the VM still works, the wait for a download is excruciating compared to broadband

To date I've examined 2718 real movies and have marked 18 of them for possible tracks. Of those, two have passed the team's first cut as WCO. A few of my others I think should be WCO as well, but the team hasn't gotten to them yet. The rest of them are probably terrestrial inclusions.

I haven't been the first one to click on anything yet, but I'll be happy to be one of the "agreements" on any real IS dust particles

The number of new S@H participants joining per day seems to be levelling off, with the total number hanging around 15,000. I'm staying at about the same level in the ranks for the 1/2 hour or so I do a day.

I wonder how many volunteers have stopped dusting?

[Tman]

O' Dave thanks for the nice comment.

I slowed down dramatically one week ago or so. I'm concerned about less news on the scanning process and the handling with the existing movies. If the actual process is that what we have here, then we could have (long ago) a "coffee break" until new movies are coming in.
It shows a pool of circa 80'000 existing Real Movies. When I take my score of ~19'500 seen Real Movies and extrapolate it with all the other searchers, I guess we have seen already all these movies sufficiently - Even yourself got the same Real Movie several times (which could be likely in this system of one (movie) pool sadly).

As far I'm not in the know where we actually stand in the process and whether seen movies will taken out or not (and if so which quantity of them remain), I not really feel like doing further in the search. Therefore I will probably leave sooner or later the "illustrious" first hundred... I guess
Apropos first hundred, I'm struck there're some searchers who be thinking about to search in each movie of the finally ~700'000 several times...

My current statistic:

Calibration Movies Answered Correctly: 7440
Calibration Movies Answered Incorrectly: 9
Your Overall Score: 7431
Total Real Movies Viewed: 19463
Your Rank: 54 out of 15138
Specificity: 99.92%
Sensitivity: 99.84%

Marked 30 of them for possible tracks (none of them as the first one too). Following have got a comment so far:

Movie 7249257V1 / 167 47 Passed cut 1 IS candidate lower right
Movie 5714462V1 / 198 77 Not Extraterrestrial inclusions
Movie 433711V1 / 251 57 Passed cut 1 another IDP
Movie 7933874V1 / 221 8 Indeterminate (but well-known by the Team)
Movie 7567884V1 / 186 45 Passed cut 1 unknown, WCO
Movie 8130472V1 / 129 45 Passed cut 1 IDP?

The still not commented rest are probably mostly terrestrial inclusions too.
If you like to have a look at, then copy and paste the movie ID in that example Movie 433711V1

[Tman]

http://stardustathome.ssl.berkeley.edu/for...opic.php?t=1186

But it wasn't the reason of lack of information - that is still misty to me.
Some clarifications though: http://stardustathome.ssl.berkeley.edu/for...&highlight=

[Tman]

The number of new S@H participants joining per day seems to be levelling off, with the total number hanging around 15,000. I'm staying at about the same level in the ranks for the 1/2 hour or so I do a day.

I wonder how many volunteers have stopped dusting?

Few days ago I started a new account in order to receive an impression of the (rear) activities there.
I was a little surprised by reaching ranks around 3400 rather soon - already after seen 500 Real Movies.
Don't know what that means exactly, but "regularly" at work were sure less than 3000 until now.