11,166 reputation
22654
bio website
location Munich, Germany
age 36
visits member for 2 years, 11 months
seen 2 hours ago

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


Dec
14
comment ReplaceImageValue and ReplacePixelValue extremely slow
If you just want to highlight pixels, you can use: HighlightImage[image, bin], which is very fast. In general, working with coordinates/indices seems to be much slower than using converting images to arrays with ImageData and using simple arithmetic.
Dec
4
comment Mask outliers in an image
You wrote "I can do the outliner detection" but you didn't give any details: What's the result of your outlier detection? A list of coordinates? A binary image (or 2d array)? Component data from ComponentMeasurements? Also: working with lists of indices is often much slower than working with a 2d array containing 0's and 1's (because of vectorization).
Nov
18
comment What does it mean: “Assuming a matrix | Use a list of lists instead”?
In Mathematica, a matrix is just a list of lists, so don't worry. This just means that the suggestions bar offers suggestions for matrices. If you clicked on "list of lists", you'd see different suggestions.
Nov
13
comment More on performance of Sum[]
I would guess that the resolution of AbsoluteTiming is system dependent. For example, on my Win 7 PC, I never get a value between 0 and 0.0005, so that's probably the resolution of some clock that AbsoluteTiming uses. If you compare only timings large compared to the clock resolution (e.g. t > 0.01), the graphs do look very similar.
Nov
13
comment Finding a specific number of eigenvalues in Mathematica
If m is a square matrix, then Eigenvalues[m,n] gives the first n eigenvalues.
Nov
13
comment FindPeaks does not smooth noise on peak?
MaxDetect[data, 0.01] where 0.01 is a "noise threshold" is also worth a try - the resulting vector is 1 for (possibly extended) maxima, and 0 elsewhere.
Nov
6
comment Best fit equation from a scanned image
Related: mathematica.stackexchange.com/questions/3831/…
Oct
29
comment Usage of @ symbol in Mathematica
Have you seen this? mathematica.stackexchange.com/questions/18393/…
Oct
25
comment how to speedup subpixel registration
The simplest way is to replace (# + {x, y}) & with TranslationTransform[{x, y}] - that makes it about 10x faster on my PC (I'm guessing ImageTransformation is optimized for TransformationFunctions). Replacing NMinimize with FindMinimum makes it 10x faster, again. You can probably make it faster if you carefully think about how ImageTransformation interpolates between pixels (there might even be a closed form solution?), but that's a lot more work for 0.2 seconds.
Sep
29
comment Curve fit based on minimal data
According to the wikipedia link, the polar form is just $r=b+a\cos\theta$ - so maybe you could convert your points to polar coordinates and use Fit?
Sep
22
comment Is there a CDF Player for Mathematica 10?
But you won't tell us where to find it?
Sep
5
comment Compose one image over another
You can use ImageApply to apply a function like Min pixel-wise. Or turn the images to normal arrays using ImageData and use MapThread.
Sep
4
comment Are there ways to create local variables analogous to a local variable in C++ and other languages?
Maybe I'm missing something. Isn't that exactly what Module does?
Aug
31
comment ImageTransformation: polar to cartesian
@paw: The black dot is there because the polar image contains no data for that location. If you use 0 in the DataRange of the inverse transform, but used 1 in the PlotRange of the polar transform, the result will be slightly skewed (by 1 pixel at the center). You can either use 0 for both, or use Padding -> "Fixed" to get rid of the black dot in the center.
Aug
12
comment How to translate and scale an image via ImageTransformation
You could just use Composition[TranslationTransform[...], ScalingTransform[...], RotationTransform[...] ] to create a complex transformation from simple ones
Jul
30
comment Automatically lining up two images that have some common elements
@Mr.Wizard: ImageTransformation can wrap. If you have per-pixel offsets (as you would have with many optical flow or non-parametric registration algorithms), this question might be useful: mathematica.stackexchange.com/questions/31274/…
Jul
25
comment Possible bug in Dynamic
@MichaelE2: You're right, this looks like a race condition: If I assign all Mathematica processes to a single CPU core, it doesn't crash anymore (or at least not that often). Maybe the WRI support people can't reproduce it because they have single core CPUs?
Jul
25
comment Possible bug in Dynamic
Yes, putting the ContourPlot inside Module[{x,y}, ... works, too. (I actually included that workaround in my WRI support ticket, but forgot it in the question here...)
Jul
23
comment Is there a way to recreate the typical Red/Blue-Postereffect using Mathematica?
I don't think you should delete the question. There are no close votes, either. But most image processing questions at least contain some img = Import[url of a good source image] code snippet, so everyone can start playing around right away, and the resulting answers are comparable.
Jul
23
comment Plot a perfect looking circle using two functions, always look like an ellipse
@TimHUK: I think the idea is that the axes in Plot often show different quantities (e.g. meters versus seconds), where PlotRange->Automatic would make no sense, and often give extremely stretched results. So WRI chose a fixed ratio that's supposed to be pretty. (But that's really just guessing.)