I have a 3D image containing distance fields, and would like to produce a mesh region from a contour at a fixed level.

I found two options for this

  1. ImageMesh[], works on binary 3D images, allows control of algorithms, marching cubes, etc.
  2. ListContourPlot3D[], extract mesh manually, no control of algorithm, cannot get a 3D region as output

I feel I must be missing a way to operate ImageMesh (or ArrayMesh) without having to turn my data into binary form, as imageData contains continuous values. I do want additional control over algorithm, which is missing from the ListContourPlot3D[] way.

Can someone suggest a better way of doing this ?

distanceFunction = SignedRegionDistance[Ball[]];
imageData = Table[distanceFunction[{x, y, z}], {x, -1, 1, 0.2}, {y, -1, 1, 0.2}, {z, -1, 1, 0.2}];
fixedLevel = 0;
imageBinarized = Binarize[Image3D[-imageData], fixedLevel]

ImageMesh[imageBinarized , Method -> #] & /@ {"Exact", "MarchingCubes", "DualMarchingCubes"}

ImageMesh output

Note that ListContourPlot3D uses original continuous data, not binary image

g = ListContourPlot3D[imageData, Contours -> {0}]

ListContour output ListContour data


How about


enter image description here

Or refine the increment in the Table command:

distanceFunction = SignedRegionDistance[Ball[]];
step = 0.02;
imageData = 
  Table[distanceFunction[{x, y, z}], {x, -1, 1, step}, {y, -1, 1, 
    step}, {z, -1, 1, step}];
fixedLevel = 0;
imageBinarized = Binarize[Image3D[-imageData], fixedLevel]

ImageMesh[imageBinarized, Method -> #] & /@ {"Exact", "MarchingCubes",

enter image description here

  • $\begingroup$ DiscretizeGraphics[g] works great, thank you! Refining the increments is not really an option, though, I hope to apply this algorithm to data that is {millions,millions, 3 to 5} in size . Having more voxels will cost me a lot. I would like to use an advanced algorithm i can control (like in ImageMesh), but with continuous data, like in ListContour... $\endgroup$ – Anatoly Oct 9 '18 at 20:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.