I'm working with Suggest Points and encountering wonderful upsides as well as gotchas/slow downs. I think this feature could save a ton of time if a particular downside could be addressed. Before I understood the "confirm to edit" note next to the greyed out suggested points/green + sign, I just added one point after the next, confident that the .0 projection errors were awesome and trustworthy. I was able to merge two burly components, but got crazy results, way out of scale one component relative to the other and radically rotated. I then checked to see where these suggested points actually lived, realized most the images were totally unrelated.
The act of searching for these images referenced by the suggested points was laborious. When you're in 6-pane view the 2DS panel is reasonably only wide enough to display 3 thumbs per row, so after noting an image number in 1DS> click 2DS> scroll scroll scroll (no page down function), mouse over thumbs to wait for rollover to display an image no.> finally get to the image> drag to a pane... all very time consuming and tedious. Additionally, or alternately, to avoid so much scrolling/hunting in 2DS I've tried the arrow keys in an active pane to navigate a sequence, appropriate to when the next suggested point is part of a sequence. I had posted this as potential bug not long ago, this behavior of not being able to access an image known to belong to a sequence, but in the 2D pane you can't get to it. Götz had suggested this might be operator error, that the accessible images in a pane are limited to an active component. I checked against that, and that's inconsistent with what I'm experiencing, which seems a bit chaotic and arbitrary when I can toggle through a sequence as expected and when I fall way short or something in between. (I appreciate any advice particular to that.)
The same holds true for using Suggest Points correctly to begin with, i.e. "confirm to edit" meaning, find the image, drag it to the pane, and confirm it relates before clicking the green + sign. Of course here we're talking potentially dozens of suggested points, so all the time you potentially save with automation setting CPs is taxed confirming so many images. I'd like to suggest improving the workflow such that "confirm to edit" is hyperlinked to call up the associated image in a 2D pane. Since you may have three images from component A on the top row and three more on the bottom row from component B, and there being new points you'll want to add in this target-rich overlap region between the two components, I can see additional value in providing a choice to either program a press/hold on "confirm to edit" to temporarily displace one image in one of the pane with the image containing the suggested point or a double click on "confirm to edit" to permanently bring up the suggested image. Since RC is usually spot on with a suggested point when it has correctly identified a related image, the temporary display of that image can be trusted with high level of confidence.
I find myself often losing time searching 2DS orienting to where a sequence lives in the pile, where I am in the sequence, where strong candidates may reside. Between various notes and views of bigger/better grid views in other image database software, I'd also like to suggest a search box function in 2DS to type in a few keys and have RC jumping to a match(s).
Please sign in to leave a comment.