8,878 reputation
22045
bio website
location Munich, Germany
age 35
visits member for 2 years, 2 months
seen 5 hours ago

I work at a small mechanical engineering company, where I develop software and image processing algorithms for camera-based inspection machines.


Apr
15
comment Why isn't Mathematica plotting my differential equations?
Have you tried evaluating e.g. {x[30.],y[30.}/.sol? This should evaluate to a pair of numbers. If it doesn't, Plot can't work, either.
Apr
14
comment Machine precision infinity
@chuy: $MaxMachineNumber is the largest finite number the machine's native floating point type can represent. But IEEE floating point numbers can also represent plus and minus infinity. I guess he means those.
Apr
12
comment Is it possible to get AltGr key to work in Math.exe?
Happens on my PC too, only in math.exe, not in other console processes. It seems to ignore num-block keys, too, even if num-lock is on. I'd say math.exe is completely unusable (interactively) with a non-us keyboard. Dumb question: Why use the console interface in the first place? Why not use the more comfortable notepad interface?
Apr
10
comment Fitting piecewise functions
I think NonlinearModelFit works my minimizing some cost function (e.g. total squared error), but the cost function can't be evaluated if your piecewise function doesn't return a value for some values of x. Have you tried using something like {(a+b)/2,A<x<B} as a fallback? Alternatively, you could come up with your own cost function and just optimize that.
Apr
9
comment Line Intensity Profile From Image
@VitaliyKaurov: I always thought an intensity profile was the sampled intensity values along a path (in this case, a line). Maybe I misunderstood the term. What would be your definition?
Feb
26
comment Correctly set up FindMinimum
I can't try without a sample image, so this is just a guess: Did you try restricting the scfac parameter to ?NumericQ (see here: mathematica.stackexchange.com/questions/18393/…). Also, you should probably Round the size parameter to ImageResize
Feb
20
comment detecting the radius of an incomplete circle
@John: I think AreaRadiusCoverage (for each connected component) takes a disc centered at the centroid, with the same area as the component, and measures the overlap between that disc and the component. For a ring, where all pixels are far from the centroid, this should always be small. Other suitable measurements might be the ConvexCoverage, or maybe something like (MaxCentroidDistance-MinCentroidDistance)/MeanCentroidDistance
Feb
20
comment detecting the radius of an incomplete circle
@Murta: It's not a duplicate (I think): The answers in that question all use the center of mass of the marked pixels (the centroid). When the circle isn't closed or the "thickness" varies, the center of mass is not the center of the best-fit circle.
Feb
4
comment Differentiating a Dot Product
You can even write D[xVec.pVec, {pVec}] to get the gradient wrt pVec, or D[xVec.pVec, {pVec, 2}] to get the Hessian. It took me a long time to figure that out...
Jan
25
comment Labelling objects in a picture with numbers
@andy: I'm sorry, but no, I won't give you my email address. I'm not a private tutor. And if I were, I would tell you the same thing: get a good book, read the documentation. It'll be quicker, more fun and you'll learn more that way than by having someone doing your homework.
Jan
7
comment Mathematica quine
I'm pretty sure #0 is cheating. It's like fopen("program.c", "r") in C.
Jan
2
comment Specific Mathematica algorithms, for example LU Decomposition
IIRC, on a x86 CPU, floating point numbers on the FPU stack are stored and processed in 80 bit long double format, even when the variable data type is 64 bit double. This can lead to the confusing effect that changing compiler options or reordering (independent) operations can change calculation results, depending on whether a value is moved to a 64-bit memory location or can stay in the stack for the whole calculation. On modern SSE-enabled CPUs, you also have separate commands that work on 64/32bit FP numbers directly. That might explain differences in the last few bits.
Dec
31
comment Why can Format be used on the left side?
I had trouble grasping "definitions" like these, too. Until I stopped thinking about them as "definitions" and started to see them as "replacement rules", i.e. telling the interpreter "if you see something that matches the pattern on the left, replace it with that stuff on the right side of the :=". Then it started making sense.
Dec
16
comment Find the best parallel lines to a given data set
@RahulNarain: Depends on the number of data points. If you have 10^6 or 10^7 points, really simple things might take some time ;-)
Dec
16
comment Find the best parallel lines to a given data set
That happens a lot. Just look up convex hull algorithms and implement one in the language of your choice. Don't try to hand in somebody else's implementation for your own. They'll know. Good luck in college!
Dec
16
comment Find the best parallel lines to a given data set
Do you realize that you posted your question on a Q&A site for the Mathematica programming environment?
Nov
11
comment Why can I still hear sound when the frequency of the periodic function inside Play is lower than the range of human ear?
@Rojo: Yes, it makes sense to me, too, now. You learn something new every day.
Nov
8
comment Why can I still hear sound when the frequency of the periodic function inside Play is lower than the range of human ear?
@DanielW: I'm curious, what's the connection between speaker size and sound wavelength? For example, I can produce a 50Hz signal with my speakers, although they're much smaller than the 6.9m wavelength.
Nov
6
comment video image recognition
@YvesKlett: On the other hand, when the players get a strong enough electric shock, they should be easier to track ;-)
Nov
6
comment video image recognition
You could try ImageFeatureTrack. But if that doesn't work, prepare to do some serious research. Attaching LEDs to their clothes when they're in the shower might be easier ;-)