Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 7
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 2412 times and has 6 replies Next Thread
Firdalmog
Cruncher
Joined: Dec 29, 2006
Post Count: 3
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Corrupted graphics?

I have been running HDC for roughly a day now, and when I took a look at the graphics I got a bit worried. Even though the workunits seem to compute just fine and submit,


this is what I see whenever I check on it. The masks look normal but the tissue sample is just a black blob with some white around the edges. So:
1) is this within normal limits?
2) if not, is it harmful to the computing?
[Dec 30, 2006 9:38:07 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Corrupted graphics?

Looking at the filter results, there is plenty of information there to process.

As I understand it, the samples are stained using different compounds. Some of these cause the image to be very dark. However, we must assume that doing this highlights features of interest to the scientists.

Of course, it is possible that some of the images will be too dark, or contain no useful information. But it's not a judgement we are in a position to make.

In this case, if you looked at the raw image, and perhaps adjusted the contrast, you would see all sorts of fine detail. The filters can detect this detail, as you can see in the filter result windows. However, the scaled down main image does lose a lot of detail - but that's only there to show us an overview.
[Dec 30, 2006 10:12:57 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Sekerob
Ace Cruncher
Joined: Jul 24, 2005
Post Count: 20043
Status: Offline
Reply to this Post  Reply with Quote 
Re: Corrupted graphics?

In the BOINC 'projects' dir u can see several large files. If u open them up in hi-resolution graphics viewer, u can see the detail and finesse, which is lost in the little circle of the HDC graphics view, but sure enough it's there. I don't have one presently, but the extension of the file to look for is pgm and is about 1.5 mb in size.

Here's a comment by one of the WCG techs of what one may expect to see.

http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=9347#74246
----------------------------------------
WCG Global & Research > Make Proposal Help: Start Here!
Please help to make the Forums an enjoyable experience for All!
[Dec 30, 2006 10:29:22 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Firdalmog
Cruncher
Joined: Dec 29, 2006
Post Count: 3
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Corrupted graphics?

Looking at the filter results, there is plenty of information there to process.

As I understand it, the samples are stained using different compounds. Some of these cause the image to be very dark. However, we must assume that doing this highlights features of interest to the scientists.

Of course, it is possible that some of the images will be too dark, or contain no useful information. But it's not a judgement we are in a position to make.

In this case, if you looked at the raw image, and perhaps adjusted the contrast, you would see all sorts of fine detail. The filters can detect this detail, as you can see in the filter result windows. However, the scaled down main image does lose a lot of detail - but that's only there to show us an overview.


If it really was because of staining, I'd agree. However, the complete lack of variance in the black makes me think that it's not a problem with contrast, but rather a one with the image display. The edges to me atleast would seem to support that idea, as they appear in a similar shade of grey-white as the properly showing slides. (Ie. the edges show up just fine, but some problem in the rendering makes it show the rest as "blank", which in this case is interpreted as black) Additionally, atleast 2 separate workunits I've gotten have had the same problem, which increases the odds of something being off with the client.

If u open them up in hi-resolution graphics viewer, u can see the detail and finesse, which is lost in the little circle of the HDC graphics view, but sure enough it's there. I don't have one presently, but the extension of the file to look for is pgm and is about 1.5 mb in size.


At the moment my client is processing a WU from Proteome Folding Phase 2, but I set my project preferences to accept only HDC work to speed up getting another WU to test this on. I'll get back to you once this happens.
[Dec 30, 2006 11:14:42 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Corrupted graphics?

Really, there's no need. There's a fair number of very dark slides, and sometimes it worries people at first.

The rough edge is just a result of the way the image is drawn, and again is nothing to worry about.

Please don't waste any more time on it -- others before you have already examined this phenomenon.
[Dec 30, 2006 12:10:42 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Firdalmog
Cruncher
Joined: Dec 29, 2006
Post Count: 3
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Corrupted graphics?

Really, there's no need. There's a fair number of very dark slides, and sometimes it worries people at first.

The rough edge is just a result of the way the image is drawn, and again is nothing to worry about.

Please don't waste any more time on it -- others before you have already examined this phenomenon.


If you insist. Here's an idea: since apparently some other people have wondered about this before, could this be added to some relevant FAQ? I did some preliminary browsing here before I created this thread but couldn't really find anything that was pertinent.
[Dec 30, 2006 1:03:14 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Alther
Former World Community Grid Tech
United States of America
Joined: Sep 30, 2004
Post Count: 414
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Corrupted graphics?

I have been running HDC for roughly a day now, and when I took a look at the graphics I got a bit worried. Even though the workunits seem to compute just fine and submit,

this is what I see whenever I check on it. The masks look normal but the tissue sample is just a black blob with some white around the edges. So:
1) is this within normal limits?
2) if not, is it harmful to the computing?

There is nothing to be worried about. As has been mentioned already, some slides are very dark, though completely black ones are rare. This is due to 3 things:

  • The amount of tissue in the slide
  • The staining.
  • The reduction in size of the actual tissue image to display on the screen. The actual image that is processed is many times larger.

The white edge is just the edge where the slide image is drawn up to. It's not actually part of the slide.

As long as the WU progresses and you see images in the distance and filter masks, it's progressing fine.
----------------------------------------
Rick Alther
Former World Community Grid Developer
[Jan 1, 2007 3:21:28 PM]   Link   Report threatening or abusive post: please login first  Go to top 
[ Jump to Last Post ]
Post new Thread