Hi Blake, Thanks for the feedback. Will contact you later as well.
Kind regards, Helmi On Thursday, January 17, 2019 at 1:25:45 AM UTC+8, Blake Dickson wrote: > > Hi Helmi, > > I would concur with Murat. There aren't any satisfying options for 3D > data. The 3D landmarks from one or more angles for each extremity is really > your only option for publications. As for your final question, yes you can > output LM 'warps' for CVA, regression, or pretty much any ordination you > want, but your options on software that can do it are limited. Stephan > Schlager's R packages are a good start as Murat already suggested and I'm > happy to help work through some approaches in R if you want to shoot me an > email. > > Cheers, > -Blake > > > Blake Dickson > PhD Candidate > Museum of Comparative Zoology > Harvard University > 26 Oxford Street > Cambridge, MA 02138 > > > On Tue, 15 Jan 2019 at 21:28, Murat Maga <ma...@uw.edu <javascript:>> > wrote: > >> Dear Helmi, >> >> >> >> Visualizing deformation for 3D shapes is somewhat difficult, especially >> if you are planning to put in a 2D medium like a paper. What you suggested >> are your basic options. You can choose to do deformation grids >> independently for each plane, that way it is a bit easier to see what’s >> going on, especially on paper. >> >> >> >> I personally like animations, that is morphing from one (or more) PC >> extreme to the other interactively as it brings out the full benefit of 3D. >> You can do the same with regression coefficients (or CVs) as well. If you >> still have access, those should be available in Landmark Editor as you >> suggest. >> >> >> >> We have a working prototype that we are hoping to release in the few 6-8 >> weeks that will let you do a bit more than what you can do in Landmark >> editor, including GPA and PCA, visualizing landmark variability, and >> plotting. I attached a screenshot to give you an idea. You can stack >> multiple PCs to visualize simultaneously. >> >> >> >> If landmark editor doesn’t work for you and you really need something >> shoot me an e-mail (off the list), I will try to help you. If you are >> comfortable with R, you can also give a Stefan Schlager’s Morpho package a >> try as well. >> >> >> >> M >> >> >> >> *From:* Helmi Hadi <helm...@gmail.com <javascript:>> >> *Sent:* Tuesday, January 15, 2019 5:30 PM >> *To:* MORPHMET <morp...@morphometrics.org <javascript:>> >> *Subject:* [MORPHMET] Different types of 2D and 3D image when >> intepreting PCA, CVA and other analysis. >> >> >> >> Dear Morphometricians, >> >> >> >> Which kind of graphical output would be better to place when interpreting >> data output? I've made an imaginary PC graph with imaginary shape for 2D >> analysis. Transformation grid/wireframe show almost the same thing, except >> that transformation grid shows the bending energy. Wireframe removes the >> bending energy, but in return you get a sort of view of the overall shape. >> Wouldn't it be better for visualization to just find the raw data point >> corresponding to the furthest ends of the PCs and just placing it there. >> Then you get the fine details (curves) of the shape as well which you >> cannot capture with landmarks, unless you are planning to fully landmark >> the object. Example here is in the Leaf morphology article by Cardini. >> <https://journals.plos.org/plosone/article/figure?id=10.1371/journal.pone.0025630.g004> >> >> It shows all three leaf morphology visual representation side by side. >> >> >> >> For 3D, it will be harder to just use wireframe/transformational grid. >> Then should I either: >> >> 1. find the raw data of the extremities of the PC1 vs PC2 graph and place >> it in the graph or >> >> 2. take a raw specimen near the midpoint 0,0 value (in PC1 vs PC2) and >> just use IDAV to morph it to whichever PC which is displayed? >> >> >> >> If the raw data output itself can be used, can it be used for CVA, or >> Regression analysis as well? You can easily display the labels in MorphoJ >> and identify which raw data it corresponds to and just place it there. >> >> >> >> Thanks >> >> >> >> Helmi >> >> -- >> MORPHMET may be accessed via its webpage at http://www.morphometrics.org >> --- >> You received this message because you are subscribed to the Google Groups >> "MORPHMET" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to morphmet+u...@morphometrics.org <javascript:>. >> >> -- >> MORPHMET may be accessed via its webpage at http://www.morphometrics.org >> --- >> You received this message because you are subscribed to the Google Groups >> "MORPHMET" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to morphmet+u...@morphometrics.org <javascript:>. >> > -- MORPHMET may be accessed via its webpage at http://www.morphometrics.org --- You received this message because you are subscribed to the Google Groups "MORPHMET" group. To unsubscribe from this group and stop receiving emails from it, send an email to morphmet+unsubscr...@morphometrics.org.
