I am trying to approach non-rigid alignment using Convex Mixed-Integer Algorithm, but I am new to computer vision. Does anyone know some sample code that is similar - so that I can use that as a reference?
The algorithm is based on this research paper:
MINA: Convex Mixed-Integer Programming for Non-Rigid Shape Alignment
Thank you for showing interest in our work.
We model the constrained optimization problem in MATLAB with the help of YALMIP library and solve it using MOSEK.
Related
I'm trying to understand how quickhull works in 3D. I already understand how the algorithm works in 2D but I just don't get how the algorithm can be implemented in 3D space. Are there any tutorials or papers that can help me?
I was suggested to create a GA algorithm to solve the problem of linear intersections identification in 2d space. I have been thinking to modify the sweep line algorithm or a similar one (like Bentley Ottman) and create a GA algorithm.
I have been studying the literature and i also came accross "Overlay of Subdivisions" problem which is quite similar, but for multiple layers of data. Not a lot information on that problem and i am struggling to understand the algorithm.
I am now thinking that in the first case a GA optimization is impossible but maybe it is possible for the overlay of subdivisions problem.
Anyone with more experience thinks that this will work or am i looking for the impossible?
I am wondering whether there are any algorithms that use the Akl-Toussaint throw-away heuristic to compute the convex hull in 3D (not just as a simple pre-processing, but as the algorithmic principle or building block). And if so, what would their expected time complexity be?
Also, I am interested in experimental comparisons of such algorithms with the more traditional algorithms in 3D (e.g., Clarkson-Shor).
I would appreciate it very much if you could point me to papers or web pages that shed some light on my questions. (Or answer them directly :-) )
I need an algorithm to re-triangulate a given 3D solid object.
Input: A triangular mesh
Output: A new triangular mesh where the triangles - as far as possible - have equal edge lengths. Some tolerance is allowed.
Currently I reflect on a clustering algorithm in order to distribute points
equally over the surface. Then I could use a 2D Delaunay triangulation and
project the triangles to the surface. But this is just a vague idea, not an
algorithm. It would be great if somebody could suggest an algorithm or a
reference to related scientific work.
CGAL provides a surface meshing algorithm that can be used for remeshing.
The documentation page is here.
You can directly try it in the demo. On windows precompiled demo is available here. You'll also need the following dll's.
References to scientific publications are available on the aforementioned page.
I have a list of (about 200-300) 2d points. I know needs to find the polygon that encloses all of them. The polygon has to be convex, and it should be as complex as possible (i.e. not a rectangular bounding box). It should find this in as low as possible time, but there are no restrictions on memory.
You may answer in pseudocode or any language you want to use.
Sounds like you're looking for a convex hull algorithm? It's been more than a decade since I was taught about these, but the name Graham Scan sticks in my mind and would probably be where I'd start.
Take a look at Graham's Algorithm.
Qhull is good software for computing 2D convex hulls.
If it is a real world problem - as in, not an academic one - there's never really a reason to solve such a generic problem yourself.