search

MINDLAB1 / SACE

The code for "Self-supervised anatomical continuity enhancement network for 7T SWI synthesis from 3T SWI"
3 stars 0 forks source link

FSL bet on 7T data #1

Open cementwise opened 1 month ago

cementwise commented 1 month ago

我在使用FSL bet去颅骨时，在3T上表现尚好，但是在7T数据上经常出现处理后图片破损（7T脑组织部分随颅骨一起被去除）的现象，调整参数并没有太大帮助，请问这个问题有什么好方法吗？

MINDLAB1 commented 1 month ago

Hi, this is an interesting question. It may be because the resolution for 7T is too high for FSL, or 7T cannot present some brain substructures. 1. If you have paired 3T-7T, do skull-stripping on 3T and then warp the mask to 7T. 2. Down-sampling the 7T and up-sampling the mask. 3. Test other tools, such as SynthiStrip.