View Full Version : Monitor color settings
2012-May-03, 04:38 AM
I am trying to learn the correct identification of craters. After completing my markup of a crowded crater image with my green rings, I got a pop-up msg to the effect of '56% score, 9 correct, 7 missed and 20 false positives - here are the correct identifications marked over yours'. The moon image now showed with 16 yellow rings. While not concerned with the score so far as for my standings, I do see it as a direct measure of my individual accuracy rating. I really do want to improve for the benefit of the project. However, not having committed all 36 of my marked locations to memory, showing just 16 yellow rings now does not help me much to determine which were the ones I missed, incorrectly marked or what the differences are between the correct marked craters and the ones I previously marked incorrectly. Since my green marks are not displayed on this corrected image and I cannot use the back function to retrieve my markups, I cannot find any way to compare my markup effort to the correct marks.
So is my monitor set wrong so something on the image is not displaying? Should I be seeing more than just the original moon image and 16 'correct' yellow ring marked craters?
For my next attempt, I got stricter on my ID's and scored (I think) 33% - with 9 correct, 0 false positive and 16 missed. The corrected screen showed 25 yellow rings but still no way to determine which were the ones where I had been correct.
2012-May-03, 05:21 AM
You can use the gallery function to compare your marks to the average and/or individual marks of other contributors (if that helps you).
2012-May-03, 01:12 PM
You must work for Staples - 'That was Easy'. Thanks, that is what was needed. I find that most deviation was not in detecting a crater but in marking the size. Some of my undetecteds were from assigning a smaller than 18 pixel size to a crater, so it did not mark for me. That and some easily correctible mistakes should boost my detection rate.
2012-May-04, 04:49 AM
Hmm. There should not be any craters in the "expert" group that are <18 px across. Are you saying that the markings on the screen from the expert list are <18 px? Or are you saying that when you look at them and re-measure, you get something like 17 px?
2012-May-05, 07:34 AM
No. My error. Gallery mostly cleared this up. Will expand on my reply below.
1) Early in my learning of this site and crater identification, I had an image where I grossly over identified numerous surface features as craters and yet missed many actual craters. Just a bit later, in trying to apply what I had been shown on the prior expert examples, I missed marking a great many actual craters. Many of the moon images only have a few craters or features to identify and it would be easy to remember your own marker locations. But these images both involved a high number of features and actual craters. I wanted to directly compare my crater ID's to the expert's ID's in the same display to better see the differences. But the expert's page only appeared after I placed all my crater markers and had indicated I was Done. 'My' markers were now gone and only the expert's markers were displayed. At this point the site did not allow me to go back and retrieve my previously marked page for re-viewing. So for any comparison of markings, I had to try to remember all the features I had either marked or not marked incorrectly while I was viewing the expert page. I had expected to find an instructional page with my marks and the expert's correct marks both displayed together for comparison and learning. When I could not find this (comparison page), I thought perhaps I had my monitor colors set wrong so that my green markers were not showing on the same page along with the yellow expert markers. So that was what I was trying to ask about. I have since determined the colors displayed on my monitor are reasonably accurate and that the site was functioning properly – it just was not setup to display what I was expecting.
The reply about using the Gallery did not give me a way to compare directly to the experts page but did provide me a way of displaying my markers in conjunction with other results to aid in improving feature identifications.
2) As a completely different discussion, I have noted that there can be significant differences in total crater counts when there are a number of craters right near the <18 to =18> pixel cut-off size. For example with a marginal (small) crater, if I expanded the cursor ring to my best estimate of the proper crater rim size, it might still be red (too small) and the crater would not mark. I am not going to change my size estimate to a larger ring just so the feature would mark. Later, when I check in the Gallery, many other participants might have judged this same marginal crater feature somewhat differently and their resulting cursor ring was just large enough to be over the minimum size. So their marker went to green and stayed to be counted. The crater size had not changed, just where the marker ring was placed by different observers judging where the crater rim was located. At one time I was seeing many more marginal craters marked as 18 or > by others that I was judging as < 18, than vice versa. Not necessarily a right or wrong thing during a learning phase, just different observation/marking tolerances resulted in different total crater counts.
3) Lastly, my reviews in the Gallery of some of my more recently completed image markings clearly showed me where I had made some very obvious errors of missing or mis-marking some craters. Since seletive Re-Do's are not an option I need to stop before getting too tired.
Powered by vBulletin® Version 4.2.0 Copyright © 2013 vBulletin Solutions, Inc. All rights reserved.