Stardust@home Suggestion Thread

[dabar]

Image Zoom 0.2.7

I find this tool of considerable use on some images... you know the ones...

Now, if we only had a CONTRAST tool...

[Nikita]

Some people have noted that the Zoom programs work great, but cause errors when clicking on a track with them. Do you get out of it before you click on a possible track? If not, are you having any errors?

[dabar]

So far, when I click on a track while zoomed all seems to work as usual.

I'd hate to think the possible track I just found with high-zoom would be ignored by the system...

[fjgiie]

Hi dabar,

There are no points awarded nor taken away when you judge a real movie, therefore you will not be able to tell if the correct coordinates were sent to the Virtual Microscope. If you click on one of those obvious calibration movies with the screen zoomed, and your score goes down, then you will know that the zoom feature interferes with sending the correct coordinates. You should be able to "go back" and see if the correct coordinates were clicked in firefox. See this link about missing calibration movies.

http://stardustathome.ssl.berkeley.edu/ ... php?t=1094

If you have a problem missing a calibration movie, zoom back to 100% before you click on any kind of track.

[scopdrvr]

Many inclusions appear at the bottom of the focus field, and a fair number of movies are "bad focus". For the areas not yet examined, and for the regions that need repeat focus movies, I have a suggestion. Might the team consider increasing the increments between each frame of the focus movie by, say, 25%? With 40 frames in the ideal movie, this would add depth of focus equal to 10 frames. Inclusions generally appear in focus for about 5 frames in depth to my measurement, so this would help us focus through and see many of the inclusions for what they are. Anything first appearing that deep would almost certainly be an inclusion and not a track. I think it would be more productive, since fewer original movies would have to be retaken for "bad focus". Since tracks are much longer, I don't think we would miss any tracks this way, either. Any comments?

[GelDelve]

Many inclusions appear at the bottom of the focus field, and a fair number of movies are "bad focus". For the areas not yet examined, and for the regions that need repeat focus movies, I have a suggestion. Might the team consider increasing the increments between each frame of the focus movie by, say, 25%? With 40 frames in the ideal movie, this would add depth of focus equal to 10 frames. Inclusions generally appear in focus for about 5 frames in depth to my measurement, so this would help us focus through and see many of the inclusions for what they are. Anything first appearing that deep would almost certainly be an inclusion and not a track. I think it would be more productive, since fewer original movies would have to be retaken for "bad focus". Since tracks are much longer, I don't think we would miss any tracks this way, either. Any comments?

I really like this idea, and would even like to see it increased anywhere from 50% to 100% (that is to say, each frame of a movie would be 1.5 to 2 times deeper than it currently is), rather than 25%, while at the same time reducing the number of frames per movie. At 50%, we might be able to look at 30 frames instead of 40. At 100%, we might be able to look at 25 frames instead of 40. Thus a 50% increase in distance at 10 fewer frames would increase the depth of view by 12.5% and decrease loading time by 25%. At a 100% increase in distance and 15 fewer frames, the depth of view increases by 25% and decreases loading time by 37.5%. Of course the trade off is the greater distance between each frame. In my experience I don't find that I gain a whole lot of knowledge looking just between two adjacent frames, but have to look at a series of at least 3 to 5 or more to make a decision, both about a track and where the surface is.

[GelDelve]

Hi dabar,

There are no points awarded nor taken away when you judge a real movie, therefore you will not be able to tell if the correct coordinates were sent to the Virtual Microscope. If you click on one of those obvious calibration movies with the screen zoomed, and your score goes down, then you will know that the zoom feature interferes with sending the correct coordinates. You should be able to "go back" and see if the correct coordinates were clicked in firefox. See this link about missing calibration movies.

http://stardustathome.ssl.berkeley.edu/ ... php?t=1094

If you have a problem missing a calibration movie, zoom back to 100% before you click on any kind of track.

Someone on this forum pointed out that if you right click the movie frame and click properties, whether in zoom mode or not, and the dimensions shown are not 375 X 500, then any click you make for a track on that movie will likely be errant. I don't know if it is true for firefox, but I do know it is true for IE.

[Titanium]

I apologize if this has already been suggested, but I really can't be buggered to read 17 pages of posts.

I keep a .rtf file with all the movies I have clicked and a small note about each (such as "probably inclusion" or "upper left"). It would be great if we could have something like this built into the "My Events" page. Just an idea I thought I would toss out there! Cheers!

[stardust1]

I think it would be great to get a high resolution view of the Phase I candidates listed in the Latest Candidates News. This could help
- to understand what were the features, scientists had looked at
- to compare and interprete the characteristics that can be found in the new 50 micron focus movies in addition to the 2 new samples added to the tutorial.
Thanks

[Koranzite]

I think that a good idea would be to implement certificates for how many movies you have viewed, not just your score.
E.g: 1000 viewed, 2500 viewed, 10000 viewed ect...

[DustSabre]

NOT a good idea; Then bad dusters who can't figure out which end is up, or a useless computer program that can't find anything, could all achieve great scores.

[Domelsmith]

Since I'm about to break into the top 1%, how about certificates for the top 1% ? Shoot, since I wouldn't be doing the work, how about the top 2%, 5%, 10% ?
Domelsmith, Temple, Texas USA

[DustSabre]

True, but the top 1% is always changing to the upside, so the certificate wouldn't be as hard to win as the same certificate later in the project.

[derekd]

Anyone agree, a horizontal scale may be better?

[DustSabre]

Personally, I like the vertical scale as it is, but maybe that's just me.