Pages: [1] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 1 post(s) |
Hinasaki Miu
School of Applied Knowledge Caldari State
0
|
Posted - 2016.05.23 21:22:09 -
[1] - Quote
Hi there,
currently I've got an accuracy at 99% and I've found that it doesn't increase anymore. Is 99% already the highest accuracy in Project Discovery?
Cheers, Hinasaki |
Tau Cabalander
Retirement Retreat Working Stiffs
6137
|
Posted - 2016.05.23 22:07:26 -
[2] - Quote
Yes. |
Hinasaki Miu
School of Applied Knowledge Caldari State
0
|
Posted - 2016.05.23 22:10:08 -
[3] - Quote
Tau Cabalander wrote:Yes. Thank you very much! |
Hug Motsu
Pandemic Horde Inc. Pandemic Horde
0
|
Posted - 2016.05.24 15:36:30 -
[4] - Quote
omg.. i just 'played' a little on this thing.. i got 48% acc.. and dropping for every try.
i guess you hafto 'learn' what is consider to be what according to the game... cus this ****....
it would be really helpful if 'cytoplasma' had a tip that say 'green dots must make up X% of the red fields to be a 'valid' choice - and stuff like this.
the UI is not user friendly. the examples pictures are so tiny and the entire info panel are just inte way when i look up details on each nucleaus/cyo/whatever
i guess i must learn how the game thinks about this - as im not blind.. i even have an visual iq closer to 200 then 100(according to weiss)
and this really pisses me off!
in short.
UI is ****.
|
Hug Motsu
Pandemic Horde Inc. Pandemic Horde
0
|
Posted - 2016.05.24 16:42:34 -
[5] - Quote
http://imgur.com/Mco9xE4
rip science. |
Mortlake
Devils Rejects 666
2127
|
Posted - 2016.05.24 16:57:58 -
[6] - Quote
I picked cytoplasm for everything and my accuracy is 99%.
I misclicked once and selected 'nucleus' when in fact it was a 'nucleus' (and I should have selected cytoplasm), hence not 100%.
Father to a murdered son. Husband to a murdered wife. Truth be told, I've been pretty unlucky.
|
Sarah Askira
South West Trading
0
|
Posted - 2016.05.24 17:50:18 -
[7] - Quote
Hug Motsu wrote:http://imgur.com/Mco9xE4
rip science.
just because it's yellow it doesnt have to be cytoskeleton. There's a reason you have the different layers. It can appear yellow if everything is green as that will also mix with the red and make it yellow at those parts. you can see multiple green spots in your image which might indicate that it's not cytoskeleton and just a lot of green all around but we can only know for sure if we see the green layer alone.
It's possible that it's a wrong sample, there are some of those but it's also possible it actually was something different and you may have seen that by looking at the green layer. |
Tau Cabalander
Retirement Retreat Working Stiffs
6141
|
Posted - 2016.05.24 18:19:31 -
[8] - Quote
Tip: Only select the BRIGHTEST green stained object.
If two or more are stained equally bright, then select all of those.
if you only select one object per slide, you will eventually hit 99%. There are few test requiring 3. |
Violet Hurst
Fedaya Recon
80
|
Posted - 2016.05.24 18:20:59 -
[9] - Quote
Sarah Askira wrote:Hug Motsu wrote:http://imgur.com/Mco9xE4
rip science. just because it's yellow it doesnt have to be cytoskeleton. There's a reason you have the different layers. It can appear yellow if everything is green as that will also mix with the red and make it yellow at those parts. you can see multiple green spots in your image which might indicate that it's not cytoskeleton and just a lot of green all around but we can only know for sure if we see the green layer alone. It's possible that it's a wrong sample, there are some of those but it's also possible it actually was something different and you may have seen that by looking at the green layer.
Umm, if I read that picture right it actually was the microtubules, whereas Hug Motsu picked intermediate fillaments (both part of the cytoskeleton, anyway).
Anyway, I found the first post here helpful: https://www.reddit.com/r/Eve/comments/42fgcf/ |
Hinasaki Miu
School of Applied Knowledge Caldari State
0
|
Posted - 2016.05.24 20:20:21 -
[10] - Quote
Hug Motsu wrote:http://imgur.com/Mco9xE4
rip science.
I've also come across some of bad samples like this. However, have to say it is relatively rare, so it doesn't affect the accuracy too much.
My experience is always check the green channel for each picture, which will help to avoid most of the mistakes. And be very careful about those pictures with clear features, but ambiguous classifications, for example a green dot near the nucleus (could be "Aggresome", "Microtubule organizing center" or "Centrosome" or even other class), since those pics are usually related to the accuracy ;-) |
|
Galaxxis
Unicorn Rampage
111
|
Posted - 2016.05.24 21:52:45 -
[11] - Quote
Hug Motsu wrote:http://imgur.com/Mco9xE4
rip science.
See how the green overlaps the red? That means it's cytoskeleton-microtubules exactly as it's marked. Maybe you should stop whining about how bad everything is and actually learn how to do it properly. You're going into it thinking you already know everything and bragging about your imagined 200 IQ, and you're just going to bumble around and **** it up. |
Maldiro Selkurk
CHEMO IMMUNO RESISTANT VIRUS X
555
|
Posted - 2016.05.25 00:19:54 -
[12] - Quote
Highest is 9 because ive proven it wont go higher than that no matter how much work at it.
Yawn,-á I'm right as usual. The predictability kinda gets boring really.
|
|
|
|
Pages: [1] :: one page |
First page | Previous page | Next page | Last page |