Brain Tumor Segmentation Model Implementation In Clara Deploy
Brain Tumor Segmentation with Clara Deploy Pipeline
We will use the installed Clara deploy SDK (refer to previous posts to install Clara Deploy in AWS) to run a brain tumor segmentation pipeline. We will use the clara ngc segmentation model for this purpose (ref - https://ngc.nvidia.com/catalog/resources/nvidia:clara:clara_ai_brain_tumor_pipeline)
Create a pipeline folder and pull the pipeline from ngc
clara pull pipeline clara_ai_brain_tumor_pipeline
Unzip the input dicom series
Unzip the model
Register the model
clara create model -p segmentation_mri_brain_tumors_br16_t1c2tc_v1/ -t tensorflow
Create the pipeline
clara create pipeline -p brain-tumor-pipeline-model-repo.yaml
Create and start the job
clara create jobs -n brain-test -p b8703c88407848a48a58496ef411daee -f ~/.clara/pipelines/clara_ai_brain_tumor_pipeline/dcm
Check the Clara Console in localhost:32002 and take the output download of the job.
View the output in 3d-slicer-segmented tumor
Location of the tumor through rendering