Player Features and Ideas Discussion

Hello! First of all, Project Discovery is actually what brought me into EVE Online, so I am a big fan of it. As soon as I started doing localization of the protein of interest I have noticed several problems:
A. Colors for channels are red/green/blue without possibility to change those, which is not a good idea for people like me (color-blind red/green, the most popular type of color-blindness). It would be much better to have a choice red/magenta so people would be able to choose what is better for them. There is a good article about making scientific images for color blind people.
B. I have no idea how "corrects" answers are being picked and that incredibly frustrates me. Let's take a look at few examples. Here are all images that I wanted to show combined, we would walk through each of them.
1. No cell-to-cell variation. There are only 5 cells (with arrows) where you can see microtubule organizing center while for other ~11 cells this center can not be observed; it might be due not the expression problem, but rather image acquisition issues, but based on image there is clearly should be cell-to-cell variation
2. Nucleous membrane selected as "correct" version. If you would look closely (2a, 2b) this is not a membrane (which should be a continuous line, or at least something like it) - it is clear that this "membrane" consist of localized spots and should not be selected as membrane
3. This is my favorite so far - all "correct" choices are wrong. I have made screenshots in a weird order, so sorry for that.
3a (bottom-right), other images; protein of interest is not in fibrillar centers of the nucleoli - for that one we would expect to observe packed spots while signal that we see is presented more like nucleus stain (if you want to distinguish protein localization in the nucleus from the cell)
3b I am not sure how to emphasize that it is not microtubules - there is almost no colocalization with microtubules + signal from the protein of interest is not localized in microtubule-like structure in the cell
3c-1-2 my picks for localization; I am not insisting that my picks are correct, but actine filament + vesicles looks more like what we see; however, this protein is weird - if we would include cell-to-cell variation then actin+golgi+TGN+vesicles might explain what we see, but none of those options are selected as "correct" answer.

At this point, not answering correctly for image like this sets off my accuracy by several percent. Taking into account that there is no way to argue about the results that seems a little bit too harsh.

Unfortunately the system does have this weak point, your answers are corrected based on the previous results. If you get a picture that was answered only once before, then you probably have a lower chance to be correct. Especially because some people are either blind, or just don't care enough to look at a picture for more than 5 seconds. And I don't think I get a percentage "refund" if everyone after me agrees with me.

I think the exoplanet research will be better. And easier for you because it will have curves instead of colors.

I honestly just stop using this because i would waste too much time trying to to find the right answer and end up being wrong 80% of the time...