Is there a well established algorithm I can steal^h^h^h^h^h copy which will draw an outline shape given an area filled with dots ?
I'm specifically thinking of this:
John Conway's Life: it might be nice way to see Life from the 'design-stance' (Dan Dennet's phrase) - and one way of doing this would be to draw outlines around either known patterns : or perhaps joining-the-dots : drawing the line on the furthest - but still interacting dots in a particular area.
So a glider would still look like a glider pretty much: but bigger shapes would just show their outline.
I haven't thought this through probably [as some patterns involve 'separated', non-interacting dots - only destined to interact in a future generation - I guess you could factor in a temporal interaction like this?] , and should have posted to 'www.halfbakery.com' , but maybe this is interesting .....
Check out Graham scan (convex hull algorithms).
Related
I'm currently working on a glsl shader (EDIT : I'm starting to think that a shader isn't necessarily the best solution and as I'm doing this in processing, I can consider a vectorial solution too) supposed to render something like this but filling the entire 2D space (or at least a larger surface):
To do so, I want to map the repeating patterns on the general leaves shapes that you can see on the top of the sketch below.
My problem is this mapping part : is it possible to find a function that project XY coordinates on the screen to another position in such a way that I can map my patterns the way I want? The leaves must have some kind of UV coordinates inside them (to be able to apply the repeating pattern) and the transformation must be a conformal map because otherwise, there would be some distortions in the pattern.
I've tried several lines of thought but I haven't managed to get the final result :
recursion :
the idea is to first cut the plane in stripes, then cut the stripes in leaves shapes that touch the top and the bottom of the stripes (because that's easier) and finally recursively cut the leaves in halves until the result looks more random. as long as the borders of the stripe aren't on the screen, it shouldn't be too noticeable. The biggest difficulty here is to avoid the distortion.
voronoi :
it may be possible to find a distance function guided by a vector field such that the Voronoi diagram looks more like what I'm looking for. However I don't think it will be possible to have the UV mapping I want. If it's the case, a good approximation woult do the trick, the result doesn't need to be exact as long as it isn't too noticable.
distortion :
it could also be possible to find a more direct way to do this projection. While desperately looking for a solution, I came across the fact that a continuous complex function is a conform map but I haven't managed to go any further.
Finaly, there may be another solution I haven't thought about and I would be glad if someone gave me a complete solution or just a new idea I haven't tried yet.
Let me explain my problem:
I have a black vector shape (let's say it's a series of joined, straight lines for now, but it'd be nice if I could also support quadratic curves).
I also have a rectangle of a predefined width and height. I'm going to place it on top of the black shape, and then take the union of the two.
My first issue is that I don't know how to quickly extract vector unions, but I think there is a well-defined formula I can figure out for myself.
My second, and more tricky issue is how to efficiently detect the position the rectangle needs to be in (i.e., what translation and rotation are needed by the matrices), in order to maximize the black, remaining after the union (see figure, below).
The red outlined shape below is ~33% black; the green is something like 85%; and there are positions for this shape & rectangle wherein either could have 100% coverage.
Obviously, I can brute-force this by trying every translation and rotation value for every point where at least part of the rectangle is touching the black shape, then keep track of the one with the most black coverage. The problem is, I can only try a finite number of positions (and may therefore miss the maximum). Apart from that, it feels very inefficient!
Can you think of a more efficient way of tackling this problem?
Something from my Uni days tells me that a Fourier transform might improve the efficiency here, but I can't figure out how I'd do that with a vector shape!
Three ideas that have promise of being faster and/or more precise than brute force search:
Suppose you have a 3d physics engine. Define a "cone-shaped" surface where the apex is at say (0,0,-1), the black polygon boundary on the z=0 plane with its centroid at the origin, and the cone surface is formed by connecting the apex with semi-infinite rays through the polygon boundary. Think of a party hat turned upside down and crumpled to the shape of the black polygon. Now constrain the rectangle to be parallel to the z=0 plane and initially so high above the cone (large z value) that it's easy to find a place where it's definitely "inside". Then let the rectangle fall downward under gravity, twisting about z and translating in x-y only as it touches the cone, staying inside all the way down until it settles and can't move any farther. The collision detection and force resolution of the physics engine takes care of the complexities. When it settles, it will be in a position of maximal coverage of the black polygon in a local sense. (If it settles with z<0, then coverage is 100%.) For the convex case it's probably a global maximum. To probabilistically improve the result for non-convex cases (like your example), you'd randomize the starting position, dropping the polygon many times, taking the best result. Note you don't really need a full blown physics engine (though they certainly exist in open source). It's enough to use collision resolution to tell you how to rotate and translate the rectangle in a pseudo-physical way as it twists and slides uniformly down the z axis as far as possible.
Different physics model. Suppose the black area is an attractive field generator in 2d following the usual inverse square rule like gravity and magnetism. Now let the rectangle drift in a damping medium responding to this field. It ought to settle with a maximal area overlapping the black area. There are problems with "nulls" like at the center of a donut, but I don't think these can ever be stable equillibria. Can they? The simulation could be easily done by modeling both shapes as particle swarms. Or since the rectangle is a simple shape and you are a physicist, you could come up with a closed form for the integral of attractive force between a point and the rectangle. This way only the black shape needs representation as particles. Come to think of it, if you can come up with a closed form for torque and linear attraction due to two triangles, then you can decompose both shapes with a (e.g. Delaunay) triangulation and get a precise answer. Unfortunately this discussion implies it can't be done analytically. So particle clouds may be the final solution. The good news is that modern processors, particularly GPUs, do very large particle computations with amazing speed. Edit: I implemented this quick and dirty. It works great for convex shapes, but concavities create stable points that aren't what you want. Using the example:
This problem is related to robot path planning. Looking at this literature may turn up some ideas In RPP you have obstacles and a robot and want to find a path the robot can travel while avoiding and/or sliding along them. If the robot is asymmetric and can rotate, then 2d planning is done in a 3d (toroidal) configuration space (C-space) where one dimension is rotation (so closes on itself). The idea is to "grow" the obstacles in C-space while shrinking the robot to a point. Growing the obstacles is achieved by computing Minkowski Differences.) If you decompose all polygons to convex shapes, then there is a simple "edge merge" algorithm for computing the MD.) When the C-space representation is complete, any 1d path that does not pierce the "grown" obstacles corresponds to continuous translation/rotation of the robot in world space that avoids the original obstacles. For your problem the white area is the obstacle and the rectangle is the robot. You're looking for any open point at all. This would correspond to 100% coverage. For the less than 100% case, the C-space would have to be a function on 3d that reflects how "bad" the intersection of the robot is with the obstacle rather than just a binary value. You're looking for the least bad point. C-space representation is an open research topic. An octree might work here.
Lots of details to think through in both cases, and they may not pan out at all, but at least these are frameworks to think more about the problem. The physics idea is a bit like using simulated spring systems to do graph layout, which has been very successful.
I don't believe it is possible to find the precise maximum for this problem, so you will need to make do with an approximation.
You could potentially render the vector image into a bitmap and use Haar features for this - they provide a very quick O(1) way of calculating the average colour of a rectangular region.
You'd still need to perform this multiple times for different rotations and positions, but it would bring it algorithmic complexity down from a naive O(n^5) to O(n^3) which may be acceptably fast. (with n here being the size of the different degrees of freedom you are scanning)
Have you thought to keep track of the remaining white space inside the blocks with something like if whitespace !== 0?
I have been looking at posts about determining if a point lies within a polygon or not and the answers are either too vague, abstract, or complex for me. So I am going to try to ask my question specific to what I need to do.
I have a set of points that describe a non-straight line (sometimes a closed polygon). I have a rectangular "view" region. I need to determine as efficiently as possible whether any of the line segments (or polygon borders) pass through the view region.
I can't simply test each point to see if it lies within the view region. It is possible for a segment to pass through the region without any point actually inside the region (ie the line is drawn across the region).
Here is an example of what I want to determine (red means the function should return true for the set of points, blue means it should return false, example uses straight lines and rectangles because I am not an artist).
Another condition I want to be able to account for (though the method/function may be a separate one), is to determine not just whether a polygon's border passes through the rectangular region, but whether the region is entirely encompassed by the polygon. The nuance here is that in the situation first described above, if I am only concerned with drawing borders, the method should return false. But in the situation described here, if I need to fill the polygon region then I need the function to return true. I currently do not need to worry about testing "donut" shaped polygons (thank God!).
Here is an example illustrating the nuance (the red rectangle does not have a single vertex or border segment passing through the on-screen region, but it should still be considered on-screen):
For the "does any line segment or polygon border pass through or lie on screen?" problem I know I can come up with a solution (albeit perhaps not an efficient one). Even though it is more verbose, the conditions are clear to me. But the second "is polygon region on screen?" problem is a little harder. I'm hoping someone might have a good suggestion for doing this. And if one solution is easily implemented on top of the other, well, booya.
As always, thank you in advance for any help or suggestions.
PS I have a function for determining line intersection, but it seems like overkill to use it to compare each segment to each side of the on-screen region because the on-screen region is ALWAYS a plain [0, 0, width, height] rectangle. Isn't there some kind of short-cut?
What you are searching for is named a Collision Detection Algorithm A Google search will lead you to plenty of implementations in various language as well as a lot of theory
There are plenty of Geometric theory behind, from the simplest bisector calculus to Constrained Delaunay Triangulations and Voronoi Diagrams (that are just examples). It depends on the shape of Object, the number of dimensions and for sure the ratio between exactness needed and computing time afforded ;-)
Good read
PS I have a function for determining
line intersection, but it seems like
overkill to use it to compare each
segment to each side of the on-screen
region because the on-screen region is
ALWAYS a plain [0, 0, width, height]
rectangle. Isn't there some kind of
short-cut?
It's not an overkill, its neccessary here. The only kind of shortcut I can think of is to hardcode values [0, 0, width, height] into that function and simplify it a bit.
Basically, I'm looking for something like this awesome research project: Gmap, which was referenced in this related SO question.
It's a rather novel data visualization that combines a network graph with an imaginary set of regions that looks like a map. Basically, the map-ification helps humans comprehend the enormous data set better.
Cool, huh? GMap doesn't appear to be open source, though I plan to contact the authors.
I already know how to create a network graph with a force-directed layout (currently using Prefuse/Flare), so an answer could be a way to layer a mapping algorithm on top of an existing graph. I'm also not concerned about the client-side at all right now - this would be a backend process, and I am flexible about technology stack and data output at this stage.
There's also this paper that describes the algorithm backing GMap. If you have heard of Voronoi diagrams (which rock, but make my head hurt), this paper is for you. I quit after Calc 1, though, so I'm hoping to avoid remembering what sigmas and epsilons are.
As a start, could you do a simple closest point sort of an algorithm? So it looks something like this: You have your force directed layout and have computed some sort of bounding box. Now you want to render it. Adjust your bounding box to line up to the origin and then as you calculate the color of each pixel, find it's closest point. This should generate some semblance of regions and should be quite simple to try out. Of course, it isn't going to be as pretty as GMap, but maybe a start? The runtime would be awful, but... I don't know about you but computing boundary lines directly sounds a lot harder to me.
I have a 3D closed mesh car object having a surface made up
triangles. I want to calculate its volume, center of volume and inertia tensor.
Could you help me
Regards.
George
For volume...
For each triangular facet, lookup its corner points. Call 'em P,Q,R.
Compute this quantity (I call it "partial volume")
pv = PxQyRz + PyQzRx + PzQxRy - PxQzRy - PyQxRz - PzQyRx
Add these together for all facets and divide by 6.
Important! The P,Q,R for each facet must be arranged clockwise as seen from outside. (Or all counter-clockwise, as long as it's consistent for all facets.)
If the mesh has any quadrilaterals, just temporarily hallucinate a diagonal joining one pair of opposite corners. That makes it into two triangles.
Practical computationial improvement: Before doing math with P,Q and R, subtract the coordinates of some "center" point C. This can be the center of mass, a midpoint between the min/max x, y and z, or any convenient point inside or near the mesh. This helps minimize truncation errors for more accurate volumes.
From numerical point of view, what you are trying to achieve is quite simple and can be reduced to calculating few quadratures. Wikipedia will provide needed information about maths behind it.
If you are looking for out-of-the-box volume calculation, take a look at this entry.
As of inertia -- shape is not enough, as you also need distribution of mass.
Well, there isn't much information on the car being provided here - you should be able to break down the car into simpler shapes - the higher degree of approximation your require - the more simpler shapes you can break it into. (This could be difficult if the car is somehow dynamically generated and completely different every time ... but I don't see that situation making any sense).
This should help with finding the Inertial Tensor of various simpler shapes: http://www.gamedev.net/community/forums/topic.asp?topic_id=57001 , finding the volumes and the likes of things like spheres and cubes is pretty common knowledge so I won't bother linking that.
I think it was Archimedes who discovered that if you submerge the car in a volume of liquid, the displaced liquid will have the same volume as the car.
I'm not sure what this would help you in this case though. Having a liquid simulation running in the background and submerging the mesh into it sounds a bit over the top. Although, I think it does work, and therefore qualifies as a (bit useless nonetheless) answer. ;^)