I'm looking for a rather fast method to decide whether an image shows more anisotropic or isotropic structures, i.e. I would like to get a kind of image analysis feature (value) which tells me something about the (an)isotropy of an image.
Usually this should be a typical task for fast Fourier transformation, I won't exclude that there are better tools but this is a typical one. Use the free ImageJ or Fiji which already contain this tool. You can test it (and perhaps some other filters). Good luck!
http://imagej.nih.gov/ij/
http://fiji.sc/Fiji
You can determined anisotropy degree by using normalised area autocorelation function and determined Str coefficient - anisotropy degree, its value is between 0 and 1. Plesase read our articles or send us yours images we can do it tougether.
Regards
1. S. Kulesza, M. Bramowicz, "A comparative study of correlation methods for determination of fractal parameters in surface characterization", Applied Surface Science, Vol.293 pp.196-201 (2014), http://dx.doi.org/10.1016/j.apsusc.2013.12.132
2. M. Bramowicz, S. Kulesza, T. Lipiński, P. Szabracki, P. Piatkowski, "Fractal Analysis of AFM Data Characterizing Strongly Isotropic and Anisotropic Surface Topography", Solid State Phenomena, Vols. 203-204 (2013)
Thank you for your reply.
Of course, FFT as well as other transformations are typical tools for the detection of a present "directionality". Do you have a specific method in mind or what method would you suggest to extract a single value for kind of quantifying (an)isotropy/directionality out of an FFT power spectrum?
At the moment I'm using a combination of pre-processing the images with anisotropic diffusion (to kind of "enhance" an existing directionality) and successively analyzing the histogram of the present gradients in the image. This works fine, but is probably slower than a method based on FFT + simple feature extraction.
Look in paper "Microstructure and Tribological Properties of FeNPs@a-C:H Films by Micromorphology Analysis and Fractal Geometry", section 2.3. Characterization of the Film Surface Texture.
Dear Mike, how are you? No idea about your question, not my field! best wishes
Giorgio
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science