Timeline for Non-algebric curve-fitting along weighted pointcloud (if possible using python)
Current License: CC BY-SA 3.0
19 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 11, 2019 at 22:06 | answer | added | P Moran | timeline score: 0 | |
| S Dec 7, 2012 at 5:31 | history | suggested | Scortchi♦ | CC BY-SA 3.0 |
fixed typos
|
| Dec 7, 2012 at 1:52 | review | Suggested edits | |||
| S Dec 7, 2012 at 5:31 | |||||
| Nov 9, 2012 at 22:07 | comment | added | whuber♦ | If the dataset is large, you might want to upload it to a repository somewhere and post a link. | |
| Nov 9, 2012 at 22:02 | comment | added | heltonbiker | @whuber Unfortunately I'll have to regenerate some data which is at work now (I'm at home now). Most probably in the next 14 hours the data will be posted here. Thanks for the interest (we could delete some of these comments thereafter). | |
| Nov 9, 2012 at 21:24 | comment | added | heltonbiker |
@whuber I'll paste a sample original surface data and some python code to render it. It is a line-structured point cloud over which I perform interpolation with scipy.interpolate.rbf (not included for now).
|
|
| Nov 9, 2012 at 20:00 | history | edited | heltonbiker | CC BY-SA 3.0 |
added 941 characters in body
|
| Nov 9, 2012 at 19:47 | comment | added | heltonbiker | @whuber I'll edit the question to clarify. There is already one answer that is working well already although it needs some tuning. I wrote another (more recent) question suggesting the use of non-parametric regression, so as to eliminate the need to consider the meaning of the data as you said, translating it into more of an abstract, geometry oriented problem (which is how it's been treated currently in my project, since there is no definite meaningful model for the back surface projection of spinal processes of the vertebrae). | |
| Nov 9, 2012 at 13:48 | vote | accept | heltonbiker | ||
| Nov 8, 2012 at 19:36 | history | edited | heltonbiker | CC BY-SA 3.0 |
added 7624 characters in body
|
| Nov 8, 2012 at 16:03 | comment | added | Bitwise | @heltonbiker what's useful about a physics-based model such as WLC is that you will get parameter values that are meaningful, for example persistence length, which quantifies the stiffness of the polymer (the spine). However I never fitted a WLC so I don't know if this is easy. You should also check that the fitting method is to suitable for noisy measurements (in case you have noise in your measurement - it wasn't really clear to me). | |
| Nov 8, 2012 at 15:33 | comment | added | heltonbiker | @Bitwise Thanks for the link, I'm gonna take a look at it! | |
| Nov 8, 2012 at 15:32 | comment | added | heltonbiker | @whuber The red line is actually a "wrong" fit given by a naive weighted spline interpolation. The color of the dots mean the weight of each point. This weight is calculated by a function which analyzes the local symmetry around the given point. Points which have very similar shapes on both sides have a high simmetry and are heuristically more likely to lie in the midline of the back surface. | |
| Nov 8, 2012 at 14:53 | comment | added | Bitwise | Have you considered using a model that would actually be related to the physics of the problem? For example, you could think of modelling the spine as a worm-like chain: en.wikipedia.org/wiki/Worm-like_chain | |
| Nov 8, 2012 at 14:33 | answer | added | Josef | timeline score: 4 | |
| Nov 8, 2012 at 14:01 | history | edited | heltonbiker | CC BY-SA 3.0 |
added 308 characters in body
|
| Nov 8, 2012 at 13:58 | comment | added | heltonbiker | @whuber Actually the points to consider are just the colored circles. The gray dots are just there to provide a reference frame of the human torso, and belong to another dataset. The colored points are stored in a matrix of N (x,y) coordinates, with shape Nx2. By differentiability I mean to be able to find the tangent (slope) of the interpolated curve for a given Y coordinate. | |
| Nov 7, 2012 at 21:14 | history | migrated | from stackoverflow.com (revisions) | ||
| Nov 6, 2012 at 17:23 | history | asked | heltonbiker | CC BY-SA 3.0 |