Replies: 2 comments
-
Hi, Sorry for the delays! I was off from work for a few weeks. The otsu method is an image thresholding strategy that finds appropriate thresholds for segragating background from main body of an image. It finds appropriate thresholds based on the histogram of the voxel/pixel intensity distribution (i.e. a valley in the histogram is a good spot to separate two different parts of the image). We use otsu thresholding as a data-driven strategy to derive an initial brain mask estimate prior to registration. This step is always run. If no_reg is selected, then no registration step is conducted, and inhomogeneity correction is in fact conducted based on the otsu-generated mask. Our script generates 4 different levels of possible otsu threshold (i.e. 4 different cuts in the histogram). So you can select how high in the histogram you want to cut the image by manipulating the threshold. 0 means no threshold is applied, so the mask generated actually covers the whole image. I hope this helps, |
Beta Was this translation helpful? Give feedback.
-
Thanks for the explanation, Gabe. This was really helpful. Thanks again. -g |
Beta Was this translation helpful? Give feedback.
-
Hello,
What is otsu and what do the integers 0,1,3,4 indicate?
Is the otsu_thresh used in conjunction with every method option? What if you select method=no_reg, would it still require an integer defined for otsu_thresh?
Thanks.
-g
Beta Was this translation helpful? Give feedback.
All reactions