Adding marks to image and other annotations?

Feature Requests
#1

Would it be reasonable to consider allowing users to mark areas in the image for reference? So we can flag problematic areas to double check as we gain better understanding of the object’s 3D structure. The marks would only be visible during that session, doesn’t necessarily have to be submitted along with the tracery after the task is completed.


In addition, sometimes I see some areas where the algorithm has apparently extended the trace too far, would it be worthwhile for users to flag that for review?
1 Like
#2

Hi alexbenjm,


That’s definitely something we may look into eventually.  I moved this thread to Feature Requests so that we can keep track of it in the future.

For the second thing that you mention, if you see a mistake in the underlying segmentation, you can abort the task and give a short message as to why.  That will flag it for us so we can take a look at it.
1 Like
#3

I second the idea of being able to mark the trace when it’s gone too far. Although I appreciate that we can’t delete it for our own protection, perhaps we could mark those areas over it in red or something?   Just a thought anyway.

1 Like
#4

Oops… just saw your reply above, Matt.  Ok… I’ll abort and explain next time. I’ve just seen a couple in a row (my first ones). Thanks

1 Like
#5

Perhaps a way of visually annotating aborted cubes would be helpful? That would make it easier to show you guys exactly what/where the problem is. It would also increase the likelihood of people explaining aborts.

1 Like
#6

Until we have the ‘abort’ option put back in you guys can post screenshots here  and I/other EyeWirers will do our best to answer any questions you might have.  We hope having an open thread for images and questions will help people learn from each other.

1 Like
#7

Add text/image to another image?Most image annotation freeware can completely get this job done.You probably could try annotation component in.NET.The marks would only be visible during that session, doesn’t necessarily
have to be submitted along with the tracery after the task is
completed.

1 Like
#8

@dragonfly please… please do not suggest any introduction of .NET code into the EyeWire codebase. EyeWire is, to my best understanding completely written in HTML5 with WebGL. as these are both relatively young standards, these guys have written most, if not all of EyeWire from scratch, ground up. So while I certainly support the ideas of Annotations for both personal reference and designating Mergers, Bleed outs, and plain old A.I. goof ups, I’m also trying to support the use of Pure WebGL/HTML5 code in this project!

  I’m not trying to put you down, but use of any .NET code as you suggested would break cross-platform compatibility!

1 Like
#9

@alexbenjm , Yes it is good idea for allowing users to mark areas in the image for reference.By this you can flag problematic areas to double check as you gain better understanding of the object’s 3D structure. You can do this annotation using classmint.com. Classmint has image annotation. For more information feel free to visit classmint.com.

1 Like
#10

It would be cool if we are working my a merger or something, if we could add the merger in in a different color, not explore mode or regular color, so we can trace around the merger and not confuse it with the real trace.

1 Like
#11

nah, just trace enough mergers until you “oh wait this is glial/merger/not good” :stuck_out_tongue: