adam submission docker

projects/Task019_ADAM/README.md

@@ -15,3 +15,7 @@ Please read the information from the homepage carefully and follow the rules and
 
 The data is now converted to the correct format and the instructions from the nnDetection README can be used to train the networks.
 
+## Submission
+The submission folder contains the scripts used for our leaderboard submissions.
+Before building the Docker Image the model directory needs to be copied to the the submission folder to be detected by the docker context. Make sure to adapt the name of the nndetection base container to the name you installed it with (/ the current nndetection version).
+Before submitting make sure to run a test prediction on the training set to double check it :)

projects/Task019_ADAM/submission/Dockerfile

+FROM nndetection:0.1
+# inference file to run predictions
+
+ARG task
+ARG model
+ARG fold
+
+ENV ENVTASK=$task ENVMODEL=$model ENVFOLD=$fold det_results=/opt/results
+ENV det_data=/opt/data
+ENV det_models=/opt/models
+
+RUN echo ${ENVTASK} ${ENVMODEL} ${ENVFOLD} \
+    && mkdir -p ${det_models}/${ENVTASK}/${ENVMODEL}/${ENVFOLD} \
+    && mkdir -p ${det_data}/${ENVTASK}/raw_splitted/imagesTs \
+    && mkdir -p ${det_results}
+
+COPY ./${ENVFOLD}/* ${det_models}/${ENVTASK}/${ENVMODEL}/${ENVFOLD}/
+COPY scripts/inference.sh .
+COPY scripts/convert.py .

projects/Task019_ADAM/submission/scripts/convert.py

+import argparse
+
+from pathlib import Path
+
+from nndet.io import load_pickle
+from nndet.core.boxes.ops_np import box_center_np
+
+THRESHOLD = 0.5
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('source', type=Path)
+    args = parser.parse_args()
+
+    source = args.source
+
+    predictions = load_pickle(source / "case_boxes.pkl")
+    boxes = predictions["pred_boxes"]
+    scores = predictions["pred_scores"]
+    keep = scores > THRESHOLD
+
+    boxes = boxes[keep]
+    if boxes.size > 0:
+        centers = box_center_np(boxes)
+    else:
+        centers = []
+
+    with open(source / "result.txt", "a") as f:
+        if len(centers) > 0:
+            for c in centers[:-1]:
+                f.write(f"{round(float(c[2]))}, {round(float(c[1]))}, {round(float(c[0]))}\n")
+            c = centers[-1]
+            f.write(f"{round(float(c[2]))}, {round(float(c[1]))}, {round(float(c[0]))}")

projects/Task019_ADAM/submission/scripts/inference.sh

+#!/bin/bash
+. /activate
+
+# bring adam file into internal structure
+cp /input/pre/struct_aligned.nii.gz ${det_data}/${ENVTASK}/raw_splitted/imagesTs/case_0000.nii.gz
+cp /input/pre/TOF.nii.gz ${det_data}/${ENVTASK}/raw_splitted/imagesTs/case_0001.nii.gz
+
+nndet_predict ${ENVTASK} ${ENVMODEL} -f -1
+cp ${det_models}/${ENVTASK}/${ENVMODEL}/${ENVFOLD}/test_predictions/* ${det_results}
+
+python convert.py ${det_results}
+cp /opt/results/result.txt /output