Reputation
20,404
Next privilege 25,000 Rep.
Access to site analytics
Badges
2 48 94
Newest
 Nice Answer
Impact
~206k people reached

Feb
8
answered Boundary identification
Feb
6
answered Is there a bug of InverseDistanceTransform?
Feb
6
comment Is there a bug of InverseDistanceTransform?
Try e.g. InverseDistanceTransform[ImageMultiply[pic, 10]] to get disks with radius 10
Feb
6
comment Is there a bug of InverseDistanceTransform?
If you evaluate Max[ImageData[pic]] with the image in the documentation, you get 13, i.e. the values of the white pixels are larger than 1, namely the disk radii in pixels. I'm guessing your image contains values between 0 and 1, so InverseDistanceTransform creates disks with radii between 0 and 1
Feb
4
answered Follow each pixel brightness in set of images (Memory constrained)
Feb
3
comment Improving visibility of text in scanned image
Related: dsp.stackexchange.com/a/1934/291
Feb
1
comment NelderMeadMinimize for 2D Gaussian fitting
Dumb question: why not just take the log of the images, then fit a 2nd order polynomial to them? That way, it's a linear least squares problem, and you can use LeastSquares
Feb
1
answered extract data from a graph image
Jan
31
comment Can I get the curvature at any point of a random curve?
@yode: You mean instead of the spanning tree? But the shortest tour (or any tour) goes through all points, then returns to the beginning. I don't want to visit all points, I want to remove those little "branches" that Thinning produces. And I don't want my path the return to the beginning. Also, finding a the shortest tour is an NP-complete problem where only approximate solutions are realistically possible, while finding a spanning tree is cheap.
Jan
31
answered Counting number of tubes in an image
Jan
30
comment Grayscale Morphology
I'm just translating the formulas on the wikipedia page azt linked. Apparently, to get the same result as in binary morphology, the kernel should contain 0 and -infinity: greyKernel = DiskMatrix[4] /. {0 -> -\[Infinity], 1 -> 0}; fil = ImageFilter[Max[# + greyKernel] &, img, 4]
Jan
29
comment Grayscale Morphology
I think (?) it should be ImageFilter[Max[# + ker] & ...
Jan
29
awarded  Good Answer
Jan
29
comment Convex and concave regions of a closed curve
Related: mathematica.stackexchange.com/q/95425/242
Jan
26
awarded  Yearling
Jan
26
comment Convert (rescale) 16 Bit png image
Why don't you just use imageData/8?
Jan
25
comment Get an array of coordinates of the edges of an image
Note that Position[ImageData[...]] returns indices, not coordinates. (The distinction is a bit confusing in MMA image processing, see for example: mathematica.stackexchange.com/a/104047/242). Use PixelValuePositions[edge, 1] to get the coordinates. Rule of thumb: If you want to manipulate the raw array returned by ImageData, use indices. If you want to draw something over the image using e.g. Show, use coordinates
Jan
24
comment How to get a full closed contour image
In coordinatescontour = [...] // DeleteSmallComponents the DeleteSmallComponents is probably a mistake. Other than that, your code seems to do what you want already.
Jan
20
answered Fitting a curve to an image object: not a superellipse, not an egg curve.. what else?
Jan
15
revised Getting image data values with coordinates in circular region
added 33 characters in body