-
Notifications
You must be signed in to change notification settings - Fork 7
Open
Description
Thanks for this wonderful segmentation model which is very useful for many applications. I have a few queries
- The model runs on CPU only even though there is GPU and take quite long (about 4 hours) to segment just thigh muscles (it breaks the input into two segments).
MuscleMap executables validated successfully.
INFO - PyTorch CUDA available: True
INFO - GPU device: NVIDIA GeForce RTX 4070 Ti SUPER
INFO - Processing subject 1
- Does this work for calf muscles segmentation? If we have the thigh and calf as separate, do we need to stitch the data together or can process independently?
- Does processing abdomen, pelvic, thigh and calf regions separately take more time or stitching them together and making it a WB scan and segmenting takes more time?
- How can I speed up the process of segmentation ?
We plan to segment a large database and each scan taking more than 12 hours for WB, makes this wonderful model not possible to use. Kindly advise.
Metadata
Metadata
Assignees
Labels
No labels