Drop mean std

README.md

@@ -113,9 +113,9 @@ Several processing steps can be submitted as a chain of Slurm jobs:
 
 ```bash
 python scripts/deploy_process.py -p cochlea-net/mobie -j <params.json> --deploy      # add image data to MoBIE, transfer to S3
-python scripts/deploy_process.py -p cochlea-net/sgn -j <params.json> --deploy        # mean_std, apply, segment SGN
-python scripts/deploy_process.py -p cochlea-net/ihc -j <params.json> --deploy        # mean_std, apply, segment IHC
-python scripts/deploy_process.py -p cochlea-net/synapses -j <params.json> --deploy   # mean_std, apply, detect synapses
+python scripts/deploy_process.py -p cochlea-net/sgn -j <params.json> --deploy        # mask, apply, segment SGN
+python scripts/deploy_process.py -p cochlea-net/ihc -j <params.json> --deploy        # mask, apply, segment IHC
+python scripts/deploy_process.py -p cochlea-net/synapses -j <params.json> --deploy   # mask, apply, detect synapses
 ```
 
 The whole chain is submitted at once.

pipelines/cochlea-net/ihc.json

@@ -1,8 +1,8 @@
 {
-    "description": "Compute the normalization, apply the IHC model and segment the prediction.",
+    "description": "Mask the volume, apply the IHC model and segment the prediction.",
     "group": "IHC",
     "steps": [
-        "mean_std_IHC",
+        "mask_IHC",
         "apply_IHC",
         "segment_IHC"
     ],

pipelines/cochlea-net/sgn.json

@@ -1,8 +1,8 @@
 {
-    "description": "Compute the normalization, apply the SGN model and segment the prediction.",
+    "description": "Mask the volume, apply the SGN model and segment the prediction.",
     "group": "SGN",
     "steps": [
-        "mean_std_SGN",
+        "mask_SGN",
         "apply_SGN",
         "segment_SGN"
     ],

pipelines/cochlea-net/synapses.json

@@ -2,7 +2,7 @@
     "description": "Detect ribbon synapses around an existing IHC segmentation and match them to it.",
     "group": "synapses",
     "steps": [
-        "mean_std_synapses",
+        "mask_synapses",
         "apply_synapses",
         "detect_synapses"
     ],

templates/cochlea-net/apply_IHC.template

@@ -38,7 +38,7 @@ echo "Output directory: ${OUTPUT_FOLDER}"
 echo "Model: ${MODEL}"
 
 # Verify the input of the job. A missing input stops the rest of the pipeline.
-for path in "$INPUT" "$OUTPUT_FOLDER"/mean_std.json "$OUTPUT_FOLDER"/mask.zarr ; do
+for path in "$INPUT" "$OUTPUT_FOLDER"/mask.zarr ; do
 	if [ ! -e "$path" ] ; then
 		echo "Missing input of the job: $path"
 		exit 1

templates/cochlea-net/apply_SGN.template

@@ -38,7 +38,7 @@ echo "Output directory: ${OUTPUT_FOLDER}"
 echo "Model: ${MODEL}"
 
 # Verify the input of the job. A missing input stops the rest of the pipeline.
-for path in "$INPUT" "$OUTPUT_FOLDER"/mean_std.json "$OUTPUT_FOLDER"/mask.zarr ; do
+for path in "$INPUT" "$OUTPUT_FOLDER"/mask.zarr ; do
 	if [ ! -e "$path" ] ; then
 		echo "Missing input of the job: $path"
 		exit 1

templates/cochlea-net/apply_synapses.template

@@ -42,8 +42,8 @@ echo "Output directory: ${OUTPUT_FOLDER}"
 echo "Model: ${MODEL}"
 
 # Verify the input of the job. A missing input stops the rest of the pipeline.
-# The mask of 'mean_std_synapses' is read by the prediction itself, not passed to it.
-for path in "$INPUT" "$OUTPUT_FOLDER"/mean_std.json "$OUTPUT_FOLDER"/mask.zarr ; do
+# The mask of 'mask_synapses' is read by the prediction itself, not passed to it.
+for path in "$INPUT" "$OUTPUT_FOLDER"/mask.zarr ; do
 	if [ ! -e "$path" ] ; then
 		echo "Missing input of the job: $path"
 		exit 1

templates/cochlea-net/detect_synapses.template

@@ -30,7 +30,7 @@ export SCRIPT_DIR=$SCRIPT_REPO/scripts
 export OUTPUT_FOLDER=<data_dir>/predictions/<cochlea>/<prediction_dir>
 
 # IHC segmentation on the S3 bucket, used to match the detections and to filter them by
-# MAX_DISTANCE. MASK_KEY is the full resolution, unlike the MASK_INPUT_KEY of 'mean_std_synapses':
+# MAX_DISTANCE. MASK_KEY is the full resolution, unlike the MASK_INPUT_KEY of 'mask_synapses':
 # the matching needs the highest resolution available.
 export MASK_PATH=<cochlea>/images/ome-zarr/<ihc_prediction>.ome.zarr
 export S3_MASK=1
@@ -39,11 +39,11 @@ export MASK_KEY="s0"
 export VOXEL_SIZE="0.38 0.38 0.38"
 export THRESHOLD=0.5
 
-# Keep MAX_DISTANCE equal to the one of 'mean_std_synapses', which checks the dilation of the mask
+# Keep MAX_DISTANCE equal to the one of 'mask_synapses', which checks the dilation of the mask
 # against it. A larger distance here matches detections the mask never covered.
 export MAX_DISTANCE=3.0
 
-# See 'mean_std_synapses' for why the MoBIE table is the local proxy of the segmentation.
+# See 'mask_synapses' for why the MoBIE table is the local proxy of the segmentation.
 export EXTERNAL_INPUT=<mobie_project>/<cochlea>/tables/<ihc_prediction>/default.tsv
 
 echo "Output directory: ${OUTPUT_FOLDER}"

templates/cochlea-net/mean_std_IHC.template → templates/cochlea-net/mask_IHC.template

@@ -1,5 +1,5 @@
 #!/bin/bash
-#SBATCH --job-name=mean-std-IHC_<cochlea_job_name>
+#SBATCH --job-name=mask-IHC_<cochlea_job_name>
 #SBATCH -t 00:30:00             # estimated time, adapt to your needs
 #SBATCH --mail-user=<user_address> # change this to your mailaddress
 #SBATCH --mail-type=FAIL        # send mail when job begins and ends

templates/cochlea-net/mean_std_SGN.template → templates/cochlea-net/mask_SGN.template

@@ -1,5 +1,5 @@
 #!/bin/bash
-#SBATCH --job-name=mean-std-SGN_<cochlea_job_name>
+#SBATCH --job-name=mask-SGN_<cochlea_job_name>
 #SBATCH -t 00:40:00             # estimated time, adapt to your needs
 #SBATCH --mail-user=<user_address> # change this to your mailaddress
 #SBATCH --mail-type=FAIL        # send mail when job begins and ends

templates/cochlea-net/mean_std_synapses.template → templates/cochlea-net/mask_synapses.template

@@ -1,5 +1,5 @@
 #!/bin/bash
-#SBATCH --job-name=mean-std-synapses_<cochlea_job_name>
+#SBATCH --job-name=mask-synapses_<cochlea_job_name>
 #SBATCH -t 00:20:00             # estimated time, adapt to your needs
 #SBATCH --mail-user=<user_address> # change this to your mailaddress
 #SBATCH --mail-type=FAIL        # send mail when job begins and ends