Update inner ear analysis

Author: constantinpape

scripts/inner_ear/training/postprocessing_and_evaluation.py

@@ -13,8 +13,8 @@
 
 from train_structure_segmentation import get_train_val_test_split
 
-# ROOT = "/home/pape/Work/data/synaptic_reconstruction/moser"
-ROOT = "/mnt/lustre-emmy-hdd/projects/nim00007/data/synaptic-reconstruction/moser"
+ROOT = "/home/pape/Work/data/synaptic_reconstruction/moser"
+# ROOT = "/mnt/lustre-emmy-hdd/projects/nim00007/data/synaptic-reconstruction/moser"
 MODEL_PATH = "/mnt/lustre-emmy-hdd/projects/nim00007/models/synaptic-reconstruction/vesicle-DA-inner_ear-v2"
 OUTPUT_ROOT = "./predictions"
 
@@ -133,7 +133,11 @@ def visualize(input_paths, name, is_nested=False, label_names=None, prefixes=Non
         for name, seg in labels.items():
             v.add_labels(seg, name=f"labels/{name}", visible=False)
         for name, seg in predictions.items():
-            v.add_labels(seg, name=name)
+            if name == "ribbon":
+                cmap = {1: "orange"}
+            else:
+                cmap = {1: "green"}
+            v.add_labels(seg, name=name, colormap=cmap)
         v.title = fname
         napari.run()
 
@@ -183,7 +187,7 @@ def segment_train_domain():
     name = "train_domain"
     run_vesicle_segmentation(paths, MODEL_PATH, name, is_nested=True)
     postprocess_structures(paths, name, is_nested=True)
-    # visualize(paths, name, is_nested=True)
+    visualize(paths, name, is_nested=True)
     results = evaluate(paths, name, is_nested=True, save_path="./results/train_domain_postprocessed.csv")
     print(results)
     print("Ribbon segmentation:", results["ribbon"].mean(), "+-", results["ribbon"].std())
@@ -229,9 +233,9 @@ def segment_rat():
 
 
 def main():
-    # segment_train_domain()
-    segment_vesicle_pools()
-    segment_rat()
+    segment_train_domain()
+    # segment_vesicle_pools()
+    # segment_rat()
 
 
 if __name__ == "__main__":

scripts/inner_ear/training/summarize_results.py

@@ -0,0 +1,116 @@
+import numpy as np
+import pandas as pd
+
+
+def summarize_source_domain():
+    result_path = "./results/train_domain_postprocessed.csv"
+    results = pd.read_csv(result_path)
+
+    ribbon_results = {
+        "dataset": [],
+        "source_model": [],
+        "target_model": [],
+    }
+    ribbon_mean = np.round(results["ribbon"].mean() * 100, 2)
+    ribbon_std = np.round(results["ribbon"].std() * 100, 2)
+    ribbon_results["dataset"].append("source")
+    ribbon_results["source_model"].append(f"{ribbon_mean} +- {ribbon_std}")
+    ribbon_results["target_model"].append("")
+    ribbon_results = pd.DataFrame(ribbon_results)
+
+    PD_results = {
+        "dataset": [],
+        "source_model": [],
+        "target_model": [],
+    }
+    PD_mean = np.round(results["PD"].mean() * 100, 2)
+    PD_std = np.round(results["PD"].std() * 100, 2)
+    PD_results["dataset"].append("source")
+    PD_results["source_model"].append(f"{PD_mean} +- {PD_std}")
+    PD_results["target_model"].append("")
+    PD_results = pd.DataFrame(PD_results)
+
+    return ribbon_results, PD_results
+
+
+def summarize_rat():
+    ribbon_results = {
+        "dataset": [],
+        "source_model": [],
+        "target_model": [],
+    }
+    PD_results = {
+        "dataset": [],
+        "source_model": [],
+        "target_model": [],
+    }
+
+    result_paths = {
+        "source_model": "results/rat_Src.csv",
+        "target_model": "results/rat_Adapted.csv",
+    }
+
+    ribbon_results["dataset"].append("source")
+    PD_results["dataset"].append("source")
+
+    for model, result_path in result_paths.items():
+        results = pd.read_csv(result_path)
+        ribbon_mean = np.round(results["ribbon"].mean() * 100, 2)
+        ribbon_std = np.round(results["ribbon"].std() * 100, 2)
+        ribbon_results[model].append(f"{ribbon_mean} +- {ribbon_std}")
+
+        PD_mean = np.round(results["PD"].mean() * 100, 2)
+        PD_std = np.round(results["PD"].std() * 100, 2)
+        PD_results[model].append(f"{PD_mean} +- {PD_std}")
+
+    ribbon_results = pd.DataFrame(ribbon_results)
+    PD_results = pd.DataFrame(PD_results)
+    return ribbon_results, PD_results
+
+
+def summarize_ves_pool():
+    ribbon_results = {
+        "dataset": [],
+        "source_model": [],
+        "target_model": [],
+    }
+    PD_results = {
+        "dataset": [],
+        "source_model": [],
+        "target_model": [],
+    }
+
+    result_paths = {
+        "source_model": "results/vesicle_pools_Src.csv",
+        "target_model": "results/vesicle_pools_Adapted.csv",
+    }
+
+    ribbon_results["dataset"].append("source")
+    PD_results["dataset"].append("source")
+
+    for model, result_path in result_paths.items():
+        results = pd.read_csv(result_path)
+        ribbon_mean = np.round(results["ribbon"].mean() * 100, 2)
+        ribbon_std = np.round(results["ribbon"].std() * 100, 2)
+        ribbon_results[model].append(f"{ribbon_mean} +- {ribbon_std}")
+
+        PD_mean = np.round(results["PD"].mean() * 100, 2)
+        PD_std = np.round(results["PD"].std() * 100, 2)
+        PD_results[model].append(f"{PD_mean} +- {PD_std}")
+
+    ribbon_results = pd.DataFrame(ribbon_results)
+    PD_results = pd.DataFrame(PD_results)
+    return ribbon_results, PD_results
+
+
+def main():
+    ribbon_results, PD_results = summarize_source_domain()
+    # ribbon_results, PD_results = summarize_ves_pool()
+    print("Ribbon")
+    print(ribbon_results)
+    print("PD")
+    print(PD_results)
+
+
+if __name__ == "__main__":
+    main()