Open thimabru1010 opened 4 months ago
The coronary_arteries model does not work as robustly as the other totalsegmentator models as it was trained only on a small dataset. It only works well if you have a image with contrast, high resolution and no artifacts.
Thanks. I am still investigating...
Does anyone have this problem? The model is outputting a model of zeros when using the task coronary_arteries. I tried using both CCAT and CT thorax images. I don´t know what can be. Does anyone have any suggestions for inspecting these input exams?