A day of problem solving pictures.
One of the things I love about using Mathematica is that it helps me better understand the nature of my programming and problem solving.
It’s a great tool to help make sense of problems, see your errors, verify and validate your computations.
Here is a set of 26 images from today’s efforts to solve a particular problem of differential geometry on densely sampled meshes.
More good bad results
Some days, things don’t go well, but you end up with interesting looking things anyway.
Glaven Explorer
Very preliminary, for Eric.
[WolframCDF source=”http://flipphillips.com/wp-content/uploads/2014/06/Glerp.cdf” width=”433″ height=”520″ altimage=”http://flipphillips.com/wp-content/uploads/2014/06/Glerp.png” altimagewidth=”433″ altimageheight=”520″]
When I do the wrong thing…
[WolframCDF source=”http://flipphillips.com/wp-content/uploads/2014/06/wtf.cdf” width=”320″ height=”415″ altimage=”https://academics.skidmore.edu/blogs/flip/?p=432″ altimagewidth=”” altimageheight=””] A small issue with a contour plot based on a very efficient Value Noise implementation (well, efficient for Mathematica).
It eventually looks like this, which is the ‘right thing’
[WolframCDF source=”http://flipphillips.com/wp-content/uploads/2014/06/notwtf.cdf” width=”320″ height=”415″ altimage=”https://academics.skidmore.edu/blogs/flip/?p=433″ altimagewidth=”” altimageheight=””]
Pegged. It will come back to you?
Doing a sort of gigantic parallel computation in Mathematica.
A little too gigantic, I’m afraid. Um- can IT deliver my new MacPro sooooon? Pleeeeease?
Abby Normal
I was trying to figure out why there were shading artifacts on a 3D object I created for some vision experiments.
I’ve got a metric-ton of Mathematica I’ve written over the years for doing computational geometry and it stood me in good stead. The blue ‘V’s you see are mismatched surface normals – they should be coincident (e.g. exactly the same) but, as you can see, there is a systematic error. Fixed it.
NewGlaven
Back in 1987 I wrote some code for generating ‘noisy objects’. I cleaned up that code over the years, enhanced some calculations, played with new Mathematica features. Now I’ve got stuff that looks like this.
I’ll put a reference to the paper on these things here soon. Sooooon.
Vision Science 2014 poster
Here’s Phillips, Mazzarella & Docter (2014) for VSS. Get the PDF here Phillips2014 (Draft of 13 May).
Specularity and shape from line drawings. If you’re in Florida next week drop in and say ‘hi’. Or not.
An interesting thought: this could be the first time Picasso and Mort Walker have been mentioned in the same work.
VSS Analysis Continues
It never ends.
Depth maps.
Pete Experiment
Going to have to hurry up to get my stuff ready for VSS.
So much data, so little time.
A project I’m working on with Julia Mazzeralla and my ol’ pal Pete Docter.