BRL-CAD provides more than two dozen types of geometry "primitives" such as ellipsoids, boxes, and cones. Every primitive is described by a collection of callback functions, for example rt_ell_bbox() returns the bounding box dimensions for an ellipsoid. Wikipedia, Wolfram Mathworld, and various other math sites (and research papers) around the web include the equations for most of our basic primitives while others are a little more tricky to compute. This task involves writing a new callback function that takes an rt_db_internal object and calculates its centroid (as a point_t 3D point). There are numerous examples in our code where we compute centroids for other primtiives. The primitives that do not already have a centroid callback are itemized in following. References:
Code:
|
| File name/URL | File size | Date submitted | |
|---|---|---|---|
| table.c | 35.8 KB | December 03 2012 07:11 UTC | |
| ebm.c | 46.1 KB | December 03 2012 07:11 UTC | |
| ebm.c | 46.1 KB | December 03 2012 07:14 UTC | |
| ebm.c.diff | 1.2 KB | December 03 2012 20:15 UTC | |
| table.c.diff | 8.8 KB | December 03 2012 20:15 UTC |
I would like to work on this task.
This task has been assigned to fernozzle. You have 72 hours to complete this task, good luck!
The work on this task is ready to be reviewed.
First, I think it would be better to submit your work as a patch, as it's rather cumbersome to have the entire .c file in front of me to look at. Minor issue though.
Looking only at your rt_eb_centroid function, I have a few issues.
First, calling your variables avgx, avgy, avgz is misleading, as "centroid" is not the same as "average". Again though, minor thing.
Second, I don't understand why you have tacked on the extra value of 0.5 to avgx and avgy. What are you doing this for?
So, very close to complete, and I'll ignore the fact that I don't know how to test it as I see Sean has mentioned fixing up analyze() as a task here, which would take care of that for me.
Cheers!
One of the mentors has sent this task back for more work. Talk to the mentor(s) assigned to this task to satisfy the requirements needed to complete this task, submit your work again and mark the task as complete once you re-submit your work.
The deadline of the task has been extended with 0 days and 12 hours.
Thanks for the suggestions.
Now, I have my reasons for doing what I did! I'm going to refer to the term "pixel" as "the rectangular prism representing the pixel"
"calling your variables avgx, avgy, avgz is misleading, as 'centroid' is not the same as 'average'"
To pluck a definition straight from Wikipedia, "informally, it is the "average" (arithmetic mean) of all points of X.", and I took this in mind when writing this function. In addition, since each pixel is an oblong (which we know the centroid of), and each pixel has exactly the same dimensions (width 1, height 1, constant tallness) (they are all represented equally, so an average of averages certainly does work here), we can represent all of the points of each pixel with a single point, its individual centroid. After averaging the centroids of all of the pixels, we can be sure that all points within this extruded bitmap have been equally represented.
"I don't understand why you have tacked on the extra value of 0.5 to avgx and avgy. What are you doing this for?"
The indices of a pixel only represent its bottom-left corner. So if I were to only average all of the indices of the pixels, that would not be averaging their centroids; that would be averaging their bottom-left corners. However, each and every pixel's centroid is exactly half-a-pixel up and half-a-pixel right from its bottom-left-corner. So, after averaging, we can simply add 0.5 to the x and y of the indices' average to reach the final centroid.
The work on this task is ready to be reviewed.
Here are some examples of the exact same centroid procedure in action (amazing balancing acts!):
http://cl.ly/image/0e2j1d0A2l2q
"If a physical object has uniform density, then its center of mass is the same as the centroid of its shape."
Melange has detected that the deadline has passed and no more work can be submitted. The submitted work should be reviewed.
Congratulations, this task has been completed successfully.
We'll likely have a follow-on task to confirm that your implementation works (maybe a unit test), but it looks reasonable at a glance. For reference, it really helps us if you create a patch file that includes all changes together (not separated per file). Running "svn diff" will give you that patch file automatically.
We'll likely have a follow-on task to confirm that your implementation works (maybe a unit test), but it looks reasonable at a glance. Thanks for your efforts and bonus points for updating the analyze command call and identifying that existing centroid function that you could leverage. ;)
sorry, ignore that last "follow-on" post. browser cache woes with the melange interface.
fernozzle, if you provide your real name, we'll include it within our authorship file to reflect your code contribution.
My real name is Michael Huang.
Thanks for the support! I really appreciate it!
femnozzle, I was apparently way too quick to accept your task... :)
Your table.c.diff file is rather incorrect (read it, did you change all of that? it says you did!) because you apparently worked on an old version and then diffed against a newer more up-to-date version of the table.c source file. If you could re-do the patch file (with both ebm.c and table.c included), it would be appreciated. You can submit it to our patches tracker on sourceforge and provide the link as a comment here:
https://sourceforge.net/tracker/?func=addgroup_id=105292atid=640804
Details on creating patches are at http://brlcad.org/wiki/Patches and at the top of http://brlcad.org/wiki/Deuces
Sorry about that. I had generated those previous patches on the locked-down Windows machines at school and hurriedly had to jump through all sorts of hoops to do so.
Now that I'm at home, however, I have submitted a proper patch (which I have read):
https://sourceforge.net/tracker/?func=detailaid=3592384group_id=105292atid=640804
Thanks!
A follow-on task has been created:
http://www.google-melange.com/gci/task/view/google/gci2012/8069204?validated
In case you didn't know, completing just one more task will get you the free Google t-shirt prize!
As GCI comes to a close, we wanted to take the time to say THANK YOU for all your efforts. This comment interface closes after GCI is over, so you're encouraged to join our mailing list where we'll be announcing contributions from GCI participants like yourelf over the upcoming months:
https://lists.sourceforge.net/lists/listinfo/brlcad-news
If you've provided your full name, we'll be sure to credit you in our authorship documentation and you'll see your name in a future announcement. If you contact us at devs@brlcad.org or via IRC, we'll even let you know when your work is integrated and follow up with updates. You're welcome and encouraged to contact us any time, especially if you have a question about how to continue participating in Open Source after GCI is over, but even if just to keep in touch. Note that ongoing participation in Open Source is one of the most impressive skills to have on your resumé. Take care, be well, and thank you again!