What are strategies I can use to fit a curve while making sure that I stay on one side of the points? For context, I'm using numpy.polyfit() to find the curve today, but this doesn't satisfy the constraint of staying "inside" a region or on one side.
If I had to take a guess, it sounds like you want your function's output values to be strictly larger or smaller than the datapoints. I would check out scipy's fitters and optimizers and especially scipy's curve_fit. A good fitting library is also lmfit.
Related
I am using Poisson surface reconstruction algorithm to reconstruct triangulated mesh surface from points. However, Poisson will always generate a watertight surface, which fills all holes with interpolation.
For some small holes that is the result of data missing, this hole filling is desirable. But for some big holes, I do not want hole filling and just want the surface to remain open.
The figure above shows my idea, the left one is a pointset with normal, the right one is reconstructed surface. I want the top of this surface remains open rather than current watertight result.
Can anyone provide me with some advice, how may I keep these big holes in Poisson surface reconstruction? Or is there any other algorithms that could solve this?
P.S.
Based on the accepted answer to this question, I understand surface reconstruction algorithms could be categorized as explicit ones and implicit ones. Poisson is implicit ones, and explicit ones could naturally handle big hole problem. But since the points data I have are mostly sparse and noisy, I would prefer an implicit one like Poisson.
Your screenshots look like you might be using MeshLab's implementation which is based on an old implementation. This implementation is not capable of trimming the surface.
The latest implementation, however, contains the SurfaceTrimmer that does exactly what you want. Take a look at the examples at the bottom of the page to see how to use it.
To use SurfaceTrimmer program, you have to first use SSDRecon program to reconstruct a mesh surface with --density, then setting a trim value would exactly remove faces under a specific threshold.
Below is a sample usage of that program on the demo eagle data
./SSDRecon --in eagle.points.ply --out eagle.screened.color.ply --depth 10 --density
./SurfaceTrimmer --in eagle.screened.color.ply --out eagle.screened.color.trimmed.ply --trim 7
How can I convert a polygon shape to a curve in JS/SVG?
I have seen this solution: http://jsdraw2d.jsfiction.com/ but this seems to be dealing with VML and not SVG.
Is there something out-of-the-box that can be used to accurately convert a polygon to a path without ANY loss of quality?
When I say path I don't mean a path with >4000 nodes. I mean a path with curves instead of many nodes. Which in turn means reducing the node count since the polygons would be converted into curves.
I assume, that while polygonizing, you sampled points on the curve, and joined them with straight lines.
The reverse process is curve fitting.
You want to do a "Hermite fitting of curve through a set of points". A little search will help you out.
There are more such fitting algorithms. This is maths based and the under the hood solution to what you want. This is also how most such problems are solved.
If you want a quick solution, you would have to find a library that does it for you. i.e take a set of points, and fits a curve through them.
Note: I assume that fitting a curve through more than 4000 nodes is going to be costly. You could try it and see the performance for yourself, as I am not sure how costly would this be. But, I would suggest that if you needed to maintain the accuracy of your boolean operation. You should not have polygonized them at first. It is just redundancy of efforts to lose accuracy only to gain it back. Boolean set operations can be be done, and are done, without polygonizing the curve data.
Links for reference, and demos
http://en.wikipedia.org/wiki/Spline_interpolation
http://www.math.ucla.edu/~baker/java/hoefer/Spline.htm
http://www.math.ucla.edu/~baker/java/hoefer/Lagrange.htm
A little background. I have a simulation that uses cubic splines for 1D trajectories. In this context, a cubic spline specifies an object's position, velocity, acceleration, and jerk as a function of time.
If you have:
initial and final values for
position, velocity, acceleration, and
time
constant-value constraints on
the maximum and minimum velocity,
acceleration, and jerk
then there is a unique spline. If you don't specify the final time, but instead want the minimum-time trajectory, then there is also a unique spline.
Actually finding these splines can be a royal pain, though. In the case where time is specified, a spline will consist of up to 7 polynomials, and the knots (transition points between polynomials) aren't known ahead of time.
This is not the usual case of fitting a spline to a set of data, it's creating splines from the boundary conditions and some additional constraints. I've read papers where people have used similar arrangements and have had similar needs, but I've never found any libraries (or even source code) that tackles generating splines of this sort. I've written some code that handles most cases, but it isn't terribly robust or fast. I'm not very worried about it being fast, but more robust would be great.
Are there any libraries that can do this available? Open source code, even if not built as a library? C, C++, Java, or Python preferred, but if it's open source other languages would still be useful as a reference.
There is a boost library for C++ that is open source and might get you half-way there.
It has all the basic building blocks you need I think (Legrendre/Laguerre/Hermite polynomials, root finding, etc...), though it comes short of actually calculating splines.
The library documentation is here so you can check for yourself: http://www.boost.org/doc/libs/1_45_0/libs/math/doc/html/index.html
The problem with splines is that you have to solve simultaneous linear equations to solve the conditions. If your situation has any more information about some of those derivatives, you may be able to use Piecewise Cubic Hermite Interpolation (PCHIP).
For example, instead of defining that jerk must be zero, you could come up with a different constraint, use PCHIP, and solve your problem greedily. Anyway, it's something to remember even if you can't use it this time.
http://www.mathworks.com/moler/interp.pdf
SciPy's interpolation functions might help... Plus you can get the derivatives or integrals of those splines easily... I'm not sure why you say "not interpolation"... It seems to me like that is what you are trying to accomplish.
Imagine we have a simple 2D drawing, filled it with lots of non-overlapping circles and only a few stars.
If we are to find all the stars among all these circles, I can think of very few methods. Brute force is one of them. Another one is possibly reduce the image size (to the optimal point where you can still distinguish the objects apart) and then apply brute force and map to the original image. The drawback of brute force is of course, it is very time consuming. I am looking for faster methods, possibly the fastest one.
What is the fastest image processing method to search for the specified item on a simple 2D image?
One typical way of looking for an object in an image is through cross correlation. Basically, you look for the position where the cross-correlation between a mask (the object you're attempting to find) and the image is the highest. That position is the likely location of the object you're trying to find.
For the sake of simplicity, I will refer to the object you're attempting to find as a star, but in general it can be any shape.
Some problems with the above approach:
The size of the mask has to match the size of the star. If you don't know the size of the star, then you will have to try different size masks. Image pyramids are more effective than just iteratively trying different size masks, but still require extra effort.
Similarly, the orientations of the mask and the star have to match. If they don't, the cross-correlation won't work.
For these reasons, the more you know about your problem, the simpler it becomes. This is the reason why people have asked you for more information in the comments. A general purpose solution doesn't really exist, to the best of my knowledge. Maybe someone more knowledgeable can correct me on this.
As you've mentioned, reducing the size of the image will help you reduce the computational time of your approach. In my opinion, it's hardly the core element of a solution -- it's just an optional optimization step.
If the shapes are easy to segment from the background, you might be able to compute distinguishing shape/color descriptors. Depending on your problem you could choose descriptors that are invariant to scale, translation or rotation (e.g. compactness, if it is unique to each shape). I do not know if this will be faster, though.
If you already know the exact shape and have an idea about the size, you might want to have a look at the Generalized Hough Transform, which is basically a formalized description of your "brute force algorithm"
As you list a property that the shapes are not overlapping then I assume an efficient algorithm would be able to
cut out all the shapes by scanning the image in some way (I can imagine relatively efficient and simple algorithm for convex shapes)
when you are left with cut out shapes you could use cross relation misha mentioned
You should describe the problem a bit better
can the shapes be rotated or scaled (or some other transform?)
is the background uniform colour
are the shapes uniform colour
are the shapes filled
Depending on the answer on the above questions you might have more less or more simple solutions.
Also, maybe this article might be interesting.
If the shapes are very regular maybe turning them into vectors could fit your needs nicely, but it might be an overkill, really depends what you want to do later.
Step 1: Thresholding - reduce the image to 1 bit (black or white) if the general image set permits it. [For the type of example you cite, my guess is thresholding would work nicely - leaving enough details to find objects].
Step 2: Optionally do some smoothing/noise removal.
Step 3: Use some clustering approach to gather the foreground objects.
Step 4: Use an appropriate heuristic to identify the objects.
The parameters in steps 1/2 will depend a lot on the type of images as well as experimentation/observation. 3 is usually straightforward if you have worked out 1/2 correctly. 4 will depend very much on the problem (for example, in your case identifying stars - which would depend on what is the actual shape of the stars expected in the images).
I want to implement the two above mentioned image resampling algorithms (bicubic and Lanczos) in C++. I know that there are dozens of existing implementations out there, but I still want to make my own. I want to make it partly because I want to understand how they work, and partly because I want to give them some capabilities not found in mainstream implementations (like configurable multi-CPU support and progress reporting).
I tried reading Wikipedia, but the stuff is a bit too dry for me. Perhaps there are some nicer explanations of these algorithms? I couldn't find anything either on SO or Google.
Added: Seems like nobody can give me a good link about these topics. Can anyone at least try to explain them here?
The basic operation principle of both algorithms is pretty simple. They're both convolution filters. A convolution filter that for each output value moves the convolution functions point of origin to be centered on the output and then multiplies all the values in the input with the value of the convolution function at that location and adds them together.
One property of convolution is that the integral of the output is the product of the integrals of the two input functions. If you consider the input and output images, then the integral means average brightness and if you want the brightness to remain the same the integral of the convolution function needs to add up to one.
One way how to understand them is to think of the convolution function as something that shows how much input pixels influence the output pixel depending on their distance.
Convolution functions are usually defined so that they are zero when the distance is larger than some value so that you don't have to consider every input value for every output value.
For lanczos interpolation the convolution function is based on the sinc(x) = sin(x*pi)/x function, but only the first few lobes are taken. Usually 3:
lanczos(x) = {
0 if abs(x) > 3,
1 if x == 0,
else sin(x*pi)/x
}
This function is called the filter kernel.
To resample with lanczos imagine you overlay the output and input over eachother, with points signifying where the pixel locations are. For each output pixel location you take a box +- 3 output pixels from that point. For every input pixel that lies in that box, calculate the value of the lanczos function at that location with the distance from the output location in output pixel coordinates as the parameter. You then need to normalize the calculated values by scaling them so that they add up to 1. After that multiply each input pixel value with the corresponding scaling value and add the results together to get the value of the output pixel.
Because lanzos function has the separability property and, if you are resizing, the grid is regular, you can optimize this by doing the convolution horizontally and vertically separately and precalculate the vertical filters for each row and horizontal filters for each column.
Bicubic convolution is basically the same, with a different filter kernel function.
To get more detail, there's a pretty good and thorough explanation in the book Digital Image Processing, section 16.3.
Also, image_operations.cc and convolver.cc in skia have a pretty well commented implementation of lanczos interpolation.
While what Ants Aasma says roughly describes the difference, I don't think it is particularly informative as to why you might do such a thing.
As far as links go, you are asking a very basic question in image processing, and any decent introductory textbook on the subject will describe this. If I remember correctly, Gonzales and Woods is decent on it, but I'm away from my books and can't check.
Now on to the particulars, it should help to think about what you are doing fundamentally. You have a square lattice of measurements that you want to interpolate new values for. In the simple case of upsampling, lets imagine you want a new measurement in between every one that you already have (e.g. double the resolution).
Now you won't get the "correct" value, because in general you don't have that information. So you have to estimate it. How to do this? A very simple way would be to linearly interpolate. Everyone knows how to do this with two points, you just draw a line between them, and read the new value off the line (in this case, at the half way point).
Now an image is two dimensional, so you really want to do this in both the left-right and up-down directions. Use the result for your estimate and voila you have "bilinear" interpolation.
The main problem with this is that it isn't very accurate, although it's better (and slower) than the "nearest neighbor" approach which is also very local and fast.
To address the first problem, you want something better than a linear fit of two points, you want to fit something to more data points (pixels), and something that can be nonlinear. A good trade off on accuracy and computational cost is something called a cubic spline. So this will give you a smooth fit line, and again you approximate your new "measurement" by the value it takes in the middle. Do this in both directions and you've got "bicubic" interpolation.
So that's more accurate, but still heavy. One way to address the speed issue is to use a convolution, which has the nice property that in the Fourier domain, it's just a multiplication, so we can implement it quite quickly. But you don't need to worry about the implementation to understand that the convolution result at any point is one function (your image) being integrated in product another, typically much smaller support (the part that is non-zero) function called the kernel), after that kernel has been centered over that particular point. In the discrete world, these are just sums of the products.
It turns out that you can design a convolution kernel that has properties quite like the cubic spline, and use that to get a fast "bicubic"
Lancsoz resampling is a similar thing, with slightly different properties in the kernel, which primarily means they will have different characteristic artifacts. You can look up the details of these kernel functions easily enough (I'm sure wikipedia has them, or any intro text). The implementations used in graphics programs tend to be highly optimized and sometimes have specialized assumptions which make them more efficient but less general.
I would like suggest the following article for a basic understanding of different image interpolation methods image interpolation via convolution. If you want to try more interpolation methods, the imageresampler is a nice open source project to begin with.
In my opinion image interpolation can be understood from two aspects, one is from function fitting perspective, and one is from convolution perspective. For example, the spline interpolation explained in image interpolation via convolution is well explained from function fitting perspective in Cubic interpolation.
Additionally, image interpolation is always related to a specific application, for example image zooming, image rotation and so on. In fact for a specific application, image interpolation can be implemented i.n a smart way. For example, image rotation can be implemented via a three-shearing method, and during each shearing operation different one-dimension interpolation algorithms can be implemented.