An Educational and Training Proposal

Archived here are older posts which are no longer relevant or were redundant.

Moderators: Stardust@home Team, DustMods

Verenique
Posts: 17
Joined: Sat May 27, 2006 3:03 am
Location: Athens, Greece
Contact:

An Educational and Training Proposal

Post by Verenique »

Hello members.

Waiting for the actual project to start I would like to share with you all an idea that just might lead us all to a more adequate level of expertise and understanding about looking for an actual track.

Let's use this or another thread and discuss each of the available movies of the tutorial session. Each one of us should comment on anything on each movie: why something looks like a track but it's not, why something doesn't look like a track but it is, questions and pitfalls.

Moderators and experts in charge of this project could comment on our views and correct us or point out strong points made by the particular member.

If this sounds like a good idea, I suggest to only have 11 topics in under each thread: one for each of the 11 available movies.

What do u thing?
Carbon number 7 inside my limbic's particular receptor, originated from a purple Supernova, generating illusions of explosions at a glance of your figure.

Verenique
Posts: 17
Joined: Sat May 27, 2006 3:03 am
Location: Athens, Greece
Contact:

View#1

Post by Verenique »

To start things out and make an example of what do I have in mind, I would like to make a few comments on View#1 of the tutorial.

What a non-expert user makes out of this movie and the comments that come with it?

First of all: one must spot the surface first of all. Make no mistake: there will be no annotations on the actual movies - real or calibration ones - and being sure about where the surface lies is of paramount importance. Why?

Because, if I got my lesson from View#1 right, surface is the place where most of the "artifacts" (--> non-dust objects and stuff) are seen gathered and focused. Being able to be pretty sure about the whereabouts of the surface helps me rule-out some of those non-dust objects. I suppose I will have to take it from there. And if what the comments say (that there will be not that much of plain dust and clutter on the actual movie's surface), this sounds as both a tricky and troublesome part of out efforts.

Second: tracks seem to share some features - kind of a hallmark:
a) they come in focus BELOW surface. (Any moderator or expert are kindly requested to correct me heavily if I'm wrong - that's the whole point behind this idea).
b) they look like lipstick: their perimeter is sharp and bold and the area they surround is luccent being as light - or even lighter than a free area of the aerogel.
c)while they come in focus below surface, you can spot their perimeter accross the slice's thickness. (what's each slice's thickness anyway?) This is in accordance to their nature: a track is actually a tunel penetrating aerogel.
d) You are suppose to see a cross-sectional area of that tunel making up a track, mostly circular. If DustTrailFinder in another thread is right, most IS dust particles hit the aerogel vertically. Of course, this is NOT to be taken as a law: View#10 shows a longitudinal section of a track - to my surprise!

And finally a question: if point c is right, why there is a certain point of focus to the tracks perimeter? I would expect more-or-less to be constantly on-focus. (I only had a year of training to microscopes during my pathology training in med-school some years ago.. :oops: )
Carbon number 7 inside my limbic's particular receptor, originated from a purple Supernova, generating illusions of explosions at a glance of your figure.

DustTrailFinder

Post by DustTrailFinder »

Each movie is made up of 40 frames and I read somewhere that they would like the first frame about 20 microns above and the last frame about 100 microns below the surface. That means each frame's focus moves down about 3 microns. In sample 1, of view#1, they label six frames as surface, which I think would be about 7.5 microns above to 7.5 microns below surface, where most of the surface features will be in focus. This movie is not really according to their ideal goal, since it appears that the last frame is about 60 microns below surface, rather than the goal of 100 and the first frame must be about 60 microns above surface instead of 20.

In sample 2 view#1, they label 10 frames as surface, which would be about plus or minus 13.5 to 14 microns above and below surface. This is most likely because the surface is not perfectly flat and the size of particles on the surface may be larger. This movie is pretty close to the ideal 20 above and 100 below. I estimate it to be about 30 microns above to 90 microns below. Sample 3 is about 50 above to 70 below, probably to focus an the aerogel fragment, since their is no trail in this movie. Same with samples 4 and 5. Sample 6 is closer to the ideal +20 to -100. If you start detecting the surface on the fifth or sixth bar from the top, then the first frame is close to 20 microns above. 8 is from about +100 to -20. Of course, 9 never even gets to the surface.

Sample 10 is the most difficult to actually pick out in my opinion. While it seems to resemble the scratch, rather than disappearing at the surface, each small segment to the left focuses as you move further down the frames. This was a rather large fragment and is still difficult to follow down. I'm thinking it'll take someone with especially keen eyes and mind to find a trail like this for the size particles we will likely encounter.

I have to agree with your observations. I add that when above and at surface, the trail is blurry and not in focus at all. In fact, it actually appears at surface to resemble a small crater. I found that after starting at the top and slowly working down the frame, it is also most helpful to then start at the bottom and find what is at focus there. Then moving toward the surface, each frame above the trail should have something in focus until relatively near the surface. Coming up from the bottom pretty much eliminates all the surface dust and imperfections that would have you searching all over the frame if you are moving down from them. Sample 7 is a good example to see how working up can be advantageous, I think.

Someone asked the question in another thread about whether the movies overlap each other. I think the answer is yes, because I tried to calculate it. It appears that the movie frames are about 139,500 square microns, about 310 X 450. Since the gel has a surface area of about 1250 square centimeters, or 125 billion square microns, with 1.6 million movies each movie would only need to cover about 218 X 358 microns, meaning that an overlap of approximately 46 microns exist on each side of the movie frames. Of course, that's still a relatively small overlap. I hope my calculations are close to accurate. Maybe one of the moderators could clue us in?

DustTrailFinder

Re: View#1

Post by DustTrailFinder »

Verenique wrote:And finally a question: if point c is right, why there is a certain point of focus to the tracks perimeter? I would expect more-or-less to be constantly on-focus. (I only had a year of training to microscopes during my pathology training in med-school some years ago.. :oops: )
This is just a guess on my part, but here goes. Layers of the aerogel closest to the point of focus will be more visible than those further away. Since the track resembles a cone, surfaces just slightly above the focus will be slightly larger and surfaces slightly below will be slightly smaller. Thus, I think the distinct perimeter is a function of the cone shaped trail. If the trail were perfectly cylindrical and straight down, then the perimeter would not be so distinct. Also, if the trail were more oblique, one side could be more distinct than the other. Maybe?

Franz
DustMod
Posts: 23
Joined: Fri May 19, 2006 6:27 pm
Location: Netherlands

Post by Franz »

My opinion is that starting the movie from above the surface of the gel is not a good choice. I prefer to start from the lowest layer. This because [taken from the tutorial movies] the trail goes so deep that it is visible [in focus] on the lowest levels. All other objects, artifacts, cracks, dust and whatever else are out of focus. Moving up to the surface the trail stays in focus [may shift position], and all the other non-trail artifacts go in/out of focus. You notice that you hit the surface when you either see the surface dust, their shadow or cracks in focus, and the trail is gone. The track does not extend above the surface.

So by going up from the bottom layer, *and* you find something in focus it may be a trail. Moving closer to the surface will introduce focus 'noise', but you know where the trail should be. On the other hand, starting from the above the surface you start with lots of objects that are in focus, and they are all most likely not to be a trail.

An analogy: you are driving your car listening to 'dance' music on the radio and hear a funny tapping noise which is not exactly in rhythm with the music. You wonder if the sound is something from your engine or from the broadcast. Is it easier to spot where it is coming from by turning the volume of the radio up (from silence), or by turning the volume down?

Verenique
Posts: 17
Joined: Sat May 27, 2006 3:03 am
Location: Athens, Greece
Contact:

Post by Verenique »

DustTrailFinder wrote: I found that after starting at the top and slowly working down the frame, it is also most helpful to then start at the bottom and find what is at focus there. Then moving toward the surface, each frame above the trail should have something in focus until relatively near the surface. Coming up from the bottom pretty much eliminates all the surface dust and imperfections that would have you searching all over the frame if you are moving down from them. Sample 7 is a good example to see how working up can be advantageous, I think.
Franz wrote:My opinion is that starting the movie from above the surface of the gel is not a good choice. I prefer to start from the lowest layer.
Excellent point from both of you!!! :D Thank you very much! Absolutely true, starting from bottom rather than surface will help us avoid confusion as of which "thing" to track down! On the other hand, however, are we absolutely sure that a dust's tarck will penetrate deep enough?

View#10 also shows a longitudinal trail. Makes me wonder.. My guess is that IS dust trail should come vertical to aerogel. Or this is not to be taken for granted?

Any moderators around?
Carbon number 7 inside my limbic's particular receptor, originated from a purple Supernova, generating illusions of explosions at a glance of your figure.

Nikita
DustMod
Posts: 994
Joined: Wed May 17, 2006 8:33 pm
Location: Indiana, USA

Post by Nikita »

I agree that a bottom up approach would be helpful! But with the possibility of angles, I will need to go up and down as I look at different parts of the screen.

By the way, the calculations you have done are great Trail Dust Finder! I found it very helpful to see what is going on.
From dust we come

joecor99
Posts: 40
Joined: Fri May 19, 2006 6:50 am
Location: Seattle, Washington
Contact:

An educational and training proposal

Post by joecor99 »

I agree with all of you that the best way of viewing the movies will be from the bottom, rather than from the top, or from the surface. I also agree with the reasoning that this will help to eliminate a lot of the clutter that will be on the surface of the aerogel. :)
And God said, Let there be light: And there was light.

Gen. 1:3

DustTrailFinder

X marks the spot

Post by DustTrailFinder »

I have a question about how we are going to mark the trails. Maybe only a mod or an alpha tester can answer it, but maybe the information is somewhere here and one or more of you have read it.

The question is: Will we mark where the trail begins at the surface, where it ends on the last frame we see the trail, on each frame from the surface to the trails end, or any frame we choose where the trail is visible? I hope we receive very good instructions about this when we are finally up and running.

If I could vote for one method or another, I would choose to mark every frame I see the trail, beginning to end. Also, I think we should only have to put a mark within the perimeter of the trail and not be scored for trying to center it. What do you guys think?

Verenique
Posts: 17
Joined: Sat May 27, 2006 3:03 am
Location: Athens, Greece
Contact:

Re: X marks the spot

Post by Verenique »

DustTrailFinder wrote:The question is: Will we mark where the trail begins at the surface, where it ends on the last frame we see the trail, on each frame from the surface to the trails end, or any frame we choose where the trail is visible? I hope we receive very good instructions about this when we are finally up and running.

If I could vote for one method or another, I would choose to mark every frame I see the trail, beginning to end. Also, I think we should only have to put a mark within the perimeter of the trail and not be scored for trying to center it. What do you guys think?
I don't even know if there will be any annotational tools available. The kind of tools - if any - made available to us will give a hint I suppose.

But I do think the answer largly depends on what us, non-specialtists of any kind think of a trail. That's why I started this thread so that any misconception could be clarified and any correct strong point stand out. But I don't see a moderator commenting any of our thoughts herein. Makes me worry to say the least...

If my vision of a trail is correct then I thing the best way to point it out is by puting an asterisk inside its perimeter for as long as you spot it, up or down the focusing. Pretty much like your suggestion.

A few things that i feel the need to clarify though: Is it equally possible to find a longitudinal trail like View#10 suggests? Or this would be a rare exception?

Should we expect a trail to penetrate enough the thickness of each slice? In those educational movies, all frames are visible for quite some distance from surface. Is it possible for a trail to be visible for only a few frames, from just below surface to 4-5 focusing lines downstream?

PS: Misconcetions of non-experts can easily render the whole project useless! Why do I have the feeling we're left alone? lol
Carbon number 7 inside my limbic's particular receptor, originated from a purple Supernova, generating illusions of explosions at a glance of your figure.

tiggertim
Posts: 47
Joined: Thu May 18, 2006 3:38 am
Location: Wanganui, New Zealand
Contact:

Post by tiggertim »

I asked the same queation a couple of weeks ago.. Under VM labeling options, I'm still waiting for a reply.

Tim
There is more to life that just the world we live on

Nikita
DustMod
Posts: 994
Joined: Wed May 17, 2006 8:33 pm
Location: Indiana, USA

Post by Nikita »

Perhaps we've had no replies because they are taking in the suggestions and seeing what would work best. I think no response may be good. We know they are watching the posts, so they must have something going on! The other possibility is that to explain it, they will need to use the films and they may want to wait to let it all come out. Kinda like trying to explain how to do something, when it's easier to show than explain.

That's my hope and I'm sticking to it. At least I hope they are working on the things we haven't gotten replies from. I'm still interested in purchasing a Stardust shirt or something, my birthday is coming up in July....
From dust we come

joecor99
Posts: 40
Joined: Fri May 19, 2006 6:50 am
Location: Seattle, Washington
Contact:

An educational and training proposal

Post by joecor99 »

I have read a post by one of the moderators that we will have to mark the tracks precisely, but there was no indication of just WHICH part of the track we should be marking. I'm as confused as the rest of you about that question. :?
And God said, Let there be light: And there was light.

Gen. 1:3

DustBuster
DustMod
Posts: 694
Joined: Wed May 17, 2006 8:12 pm
Location: Horsetown, USA

Post by DustBuster »

If you go back to the tutorial movies and move your pointer over the images, you should notice a link in the info bar at the bottom of your browser that is tracking the coordinates of your pointer as you move it across the image.
The top-left of the image coordinates are 0,0 and the bottom right coordinates are 499,374- giving us the image size of 500X375 pixels. That, coupled with the fact that each image has a unique name, will give the reviewer(s) the precise image frame and location of the pixel where you clicked. If you are within 20 or so pixels of the track, it should be easily identified.

Wolter
DustMod
Posts: 457
Joined: Mon May 22, 2006 2:23 am
Location: Enkhuizen, the Netherlands

Post by Wolter »

During alpha testing you could only mark 1 position for the track. You simply move the mouse pointer to where you think the track is, on the most visible frame, and click on it. The position where you have clicked is registered, the clip is marked as processed and a new clip will download to your PC.
So it is a one click process. If there is no track or when the clip is out of focus you have to click on the appropiate button to process the clip and continue.
Just dusting... Image

Post Reply