From 0ec513840938fc36c2bdd5dd3af628e4202cdc57 Mon Sep 17 00:00:00 2001
From: jungfrau <jungfrau@mx-spark-01.psi.ch>
Date: Tue, 29 Sep 2026 18:00:49 +0200
Subject: [PATCH] all models re-measured on the 203-frame gold set; comparison
 on the unseen frames updated

---
 .../gold_report.json                          | 109 ++-
 .../gold_report.json                          | 109 ++-
 .../gold_report.json                          | 109 ++-
 .../gold_report.json                          | 105 ++-
 .../gold_report.json                          | 105 ++-
 .../gold_report.json                          |   2 +-
 .../gold_report.json                          | 102 +--
 .../gold_report.json                          |  15 +-
 runs/clean_gold_comparison.json               | 660 +++++++++---------
 runs/clean_gold_comparison.txt                |  16 +-
 10 files changed, 634 insertions(+), 698 deletions(-)

diff --git a/models/registry/yolo26n-seg-alc-v1-2026-09-28/gold_report.json b/models/registry/yolo26n-seg-alc-v1-2026-09-28/gold_report.json
index 4d02720..d0658e6 100644
--- a/models/registry/yolo26n-seg-alc-v1-2026-09-28/gold_report.json
+++ b/models/registry/yolo26n-seg-alc-v1-2026-09-28/gold_report.json
@@ -1,52 +1,52 @@
 {
   "per_class": {
     "loop_all": {
-      "tp": 110,
-      "fp": 11,
-      "fn": 7,
-      "precision": 0.9090909090909091,
-      "recall": 0.9401709401709402,
-      "f1": 0.9243697478991597,
-      "mean_mask_iou": 0.8830125353939965,
-      "n_matched": 110
+      "tp": 180,
+      "fp": 19,
+      "fn": 12,
+      "precision": 0.9045226130653267,
+      "recall": 0.9375,
+      "f1": 0.9207161125319694,
+      "mean_mask_iou": 0.883077126387716,
+      "n_matched": 180
     },
     "pin": {
-      "tp": 85,
-      "fp": 11,
-      "fn": 4,
-      "precision": 0.8854166666666666,
-      "recall": 0.9550561797752809,
-      "f1": 0.918918918918919,
-      "mean_mask_iou": 0.9207077103499136,
-      "n_matched": 85
+      "tp": 136,
+      "fp": 19,
+      "fn": 10,
+      "precision": 0.8774193548387097,
+      "recall": 0.9315068493150684,
+      "f1": 0.9036544850498338,
+      "mean_mask_iou": 0.9171711801294601,
+      "n_matched": 136
     },
     "crystal": {
-      "tp": 20,
-      "fp": 13,
-      "fn": 13,
-      "precision": 0.6060606060606061,
-      "recall": 0.6060606060606061,
-      "f1": 0.6060606060606061,
-      "mean_mask_iou": 0.7936162087198457,
-      "n_matched": 20
+      "tp": 36,
+      "fp": 28,
+      "fn": 30,
+      "precision": 0.5625,
+      "recall": 0.5454545454545454,
+      "f1": 0.5538461538461538,
+      "mean_mask_iou": 0.7720071781000309,
+      "n_matched": 36
     },
     "loop_face": {
-      "tp": 85,
-      "fp": 14,
-      "fn": 20,
-      "precision": 0.8585858585858586,
-      "recall": 0.8095238095238095,
-      "f1": 0.8333333333333333,
-      "mean_mask_iou": 0.8623671360009486,
-      "n_matched": 85
+      "tp": 140,
+      "fp": 29,
+      "fn": 32,
+      "precision": 0.8284023668639053,
+      "recall": 0.813953488372093,
+      "f1": 0.8211143695014664,
+      "mean_mask_iou": 0.8621656147862643,
+      "n_matched": 140
     },
     "ice": {
       "tp": 4,
-      "fp": 11,
-      "fn": 28,
-      "precision": 0.26666666666666666,
-      "recall": 0.125,
-      "f1": 0.1702127659574468,
+      "fp": 13,
+      "fn": 36,
+      "precision": 0.23529411764705882,
+      "recall": 0.1,
+      "f1": 0.14035087719298245,
       "mean_mask_iou": 0.8004176972482764,
       "n_matched": 4
     },
@@ -62,32 +62,21 @@
     }
   },
   "overall": {
-    "tp": 304,
-    "fp": 60,
-    "fn": 73,
-    "precision": 0.8351648351648352,
-    "recall": 0.8063660477453581,
-    "f1": 0.8205128205128206,
-    "mean_mask_iou": 0.8808116309097135,
-    "n_matched": 304
+    "tp": 496,
+    "fp": 108,
+    "fn": 121,
+    "precision": 0.8211920529801324,
+    "recall": 0.8038897893030794,
+    "f1": 0.8124488124488125,
+    "mean_mask_iou": 0.8777949163670699,
+    "n_matched": 496
   },
   "dataset": {
-    "gold_images": 120,
-    "pred_images": 120,
+    "gold_images": 203,
+    "pred_images": 203,
     "missing_predictions": 0,
     "iou_threshold": 0.5
   },
-  "gate": {
-    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-overlap-false-2026-04-12/gold_report.json",
-    "f1_delta": 0.0452,
-    "mask_iou_delta": 0.0327,
-    "mask_score": 0.7103,
-    "mask_score_baseline": 0.6775,
-    "matched_only_iou_delta": -0.0217,
-    "class_f1_delta": {
-      "loop_all": 0.0786,
-      "pin": 0.0339
-    },
-    "pass": true
-  }
+  "weights": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-2026-09-28/weights.pt",
+  "measured_at": "2026-09-29T17:58:52"
 }
diff --git a/models/registry/yolo26n-seg-alc-v1-800px-2026-09-29/gold_report.json b/models/registry/yolo26n-seg-alc-v1-800px-2026-09-29/gold_report.json
index 53beb71..1467ee8 100644
--- a/models/registry/yolo26n-seg-alc-v1-800px-2026-09-29/gold_report.json
+++ b/models/registry/yolo26n-seg-alc-v1-800px-2026-09-29/gold_report.json
@@ -1,52 +1,52 @@
 {
   "per_class": {
     "loop_all": {
-      "tp": 112,
-      "fp": 15,
-      "fn": 5,
-      "precision": 0.8818897637795275,
-      "recall": 0.9572649572649573,
-      "f1": 0.9180327868852459,
-      "mean_mask_iou": 0.8941201653150674,
-      "n_matched": 112
+      "tp": 182,
+      "fp": 28,
+      "fn": 10,
+      "precision": 0.8666666666666667,
+      "recall": 0.9479166666666666,
+      "f1": 0.9054726368159204,
+      "mean_mask_iou": 0.8922214036056925,
+      "n_matched": 182
     },
     "pin": {
-      "tp": 85,
-      "fp": 8,
-      "fn": 4,
-      "precision": 0.9139784946236559,
-      "recall": 0.9550561797752809,
-      "f1": 0.934065934065934,
-      "mean_mask_iou": 0.9246052204787694,
-      "n_matched": 85
+      "tp": 132,
+      "fp": 18,
+      "fn": 14,
+      "precision": 0.88,
+      "recall": 0.9041095890410958,
+      "f1": 0.8918918918918919,
+      "mean_mask_iou": 0.9209259825560847,
+      "n_matched": 132
     },
     "crystal": {
-      "tp": 21,
-      "fp": 15,
-      "fn": 12,
-      "precision": 0.5833333333333334,
-      "recall": 0.6363636363636364,
-      "f1": 0.6086956521739131,
-      "mean_mask_iou": 0.8332041250658937,
-      "n_matched": 21
+      "tp": 36,
+      "fp": 37,
+      "fn": 30,
+      "precision": 0.4931506849315068,
+      "recall": 0.5454545454545454,
+      "f1": 0.5179856115107914,
+      "mean_mask_iou": 0.8104959453139609,
+      "n_matched": 36
     },
     "loop_face": {
-      "tp": 74,
-      "fp": 17,
-      "fn": 31,
-      "precision": 0.8131868131868132,
-      "recall": 0.7047619047619048,
-      "f1": 0.7551020408163266,
-      "mean_mask_iou": 0.8798695876272729,
-      "n_matched": 74
+      "tp": 130,
+      "fp": 29,
+      "fn": 42,
+      "precision": 0.8176100628930818,
+      "recall": 0.7558139534883721,
+      "f1": 0.785498489425982,
+      "mean_mask_iou": 0.8756401987234087,
+      "n_matched": 130
     },
     "ice": {
       "tp": 3,
-      "fp": 3,
-      "fn": 29,
-      "precision": 0.5,
-      "recall": 0.09375,
-      "f1": 0.15789473684210525,
+      "fp": 4,
+      "fn": 37,
+      "precision": 0.42857142857142855,
+      "recall": 0.075,
+      "f1": 0.1276595744680851,
       "mean_mask_iou": 0.9028650216500921,
       "n_matched": 3
     },
@@ -62,32 +62,21 @@
     }
   },
   "overall": {
-    "tp": 295,
-    "fp": 58,
-    "fn": 82,
-    "precision": 0.8356940509915014,
-    "recall": 0.7824933687002652,
-    "f1": 0.8082191780821918,
-    "mean_mask_iou": 0.8950818082431701,
-    "n_matched": 295
+    "tp": 483,
+    "fp": 116,
+    "fn": 134,
+    "precision": 0.8063439065108514,
+    "recall": 0.7828200972447326,
+    "f1": 0.794407894736842,
+    "mean_mask_iou": 0.8895780540040067,
+    "n_matched": 483
   },
   "dataset": {
-    "gold_images": 120,
-    "pred_images": 120,
+    "gold_images": 203,
+    "pred_images": 203,
     "missing_predictions": 0,
     "iou_threshold": 0.5
   },
-  "gate": {
-    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-2026-09-28/gold_report.json",
-    "f1_delta": -0.0123,
-    "mask_iou_delta": -0.0099,
-    "mask_score": 0.7004,
-    "mask_score_baseline": 0.7103,
-    "matched_only_iou_delta": 0.0143,
-    "class_f1_delta": {
-      "loop_all": -0.0063,
-      "pin": 0.0151
-    },
-    "pass": false
-  }
+  "weights": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-800px-2026-09-29/weights.pt",
+  "measured_at": "2026-09-29T17:58:59"
 }
diff --git a/models/registry/yolo26n-seg-alc-v1-humanonly-2026-09-28/gold_report.json b/models/registry/yolo26n-seg-alc-v1-humanonly-2026-09-28/gold_report.json
index 83c6b64..e7c3c32 100644
--- a/models/registry/yolo26n-seg-alc-v1-humanonly-2026-09-28/gold_report.json
+++ b/models/registry/yolo26n-seg-alc-v1-humanonly-2026-09-28/gold_report.json
@@ -1,52 +1,52 @@
 {
   "per_class": {
     "loop_all": {
-      "tp": 113,
-      "fp": 7,
-      "fn": 4,
-      "precision": 0.9416666666666667,
-      "recall": 0.9658119658119658,
-      "f1": 0.9535864978902954,
-      "mean_mask_iou": 0.8770614850251075,
-      "n_matched": 113
+      "tp": 181,
+      "fp": 16,
+      "fn": 11,
+      "precision": 0.9187817258883249,
+      "recall": 0.9427083333333334,
+      "f1": 0.9305912596401028,
+      "mean_mask_iou": 0.8815536331285515,
+      "n_matched": 181
     },
     "pin": {
-      "tp": 84,
-      "fp": 9,
-      "fn": 5,
-      "precision": 0.9032258064516129,
-      "recall": 0.9438202247191011,
-      "f1": 0.9230769230769231,
-      "mean_mask_iou": 0.9258323755269584,
-      "n_matched": 84
+      "tp": 135,
+      "fp": 14,
+      "fn": 11,
+      "precision": 0.9060402684563759,
+      "recall": 0.9246575342465754,
+      "f1": 0.9152542372881356,
+      "mean_mask_iou": 0.9103606709842429,
+      "n_matched": 135
     },
     "crystal": {
-      "tp": 21,
-      "fp": 24,
-      "fn": 12,
-      "precision": 0.4666666666666667,
-      "recall": 0.6363636363636364,
-      "f1": 0.5384615384615385,
-      "mean_mask_iou": 0.7929802397975723,
-      "n_matched": 21
+      "tp": 39,
+      "fp": 40,
+      "fn": 27,
+      "precision": 0.4936708860759494,
+      "recall": 0.5909090909090909,
+      "f1": 0.5379310344827586,
+      "mean_mask_iou": 0.7619988195646054,
+      "n_matched": 39
     },
     "loop_face": {
-      "tp": 87,
-      "fp": 17,
-      "fn": 18,
-      "precision": 0.8365384615384616,
-      "recall": 0.8285714285714286,
-      "f1": 0.8325358851674641,
-      "mean_mask_iou": 0.8578543451465753,
-      "n_matched": 87
+      "tp": 139,
+      "fp": 31,
+      "fn": 33,
+      "precision": 0.8176470588235294,
+      "recall": 0.8081395348837209,
+      "f1": 0.8128654970760234,
+      "mean_mask_iou": 0.8641243122696176,
+      "n_matched": 139
     },
     "ice": {
       "tp": 5,
-      "fp": 12,
-      "fn": 27,
-      "precision": 0.29411764705882354,
-      "recall": 0.15625,
-      "f1": 0.20408163265306123,
+      "fp": 13,
+      "fn": 35,
+      "precision": 0.2777777777777778,
+      "recall": 0.125,
+      "f1": 0.1724137931034483,
       "mean_mask_iou": 0.7708521231731791,
       "n_matched": 5
     },
@@ -62,32 +62,21 @@
     }
   },
   "overall": {
-    "tp": 310,
-    "fp": 69,
-    "fn": 67,
-    "precision": 0.8179419525065963,
-    "recall": 0.8222811671087533,
-    "f1": 0.8201058201058201,
-    "mean_mask_iou": 0.8774775517144151,
-    "n_matched": 310
+    "tp": 499,
+    "fp": 114,
+    "fn": 118,
+    "precision": 0.8140293637846656,
+    "recall": 0.8087520259319287,
+    "f1": 0.8113821138211383,
+    "mean_mask_iou": 0.8740388620511081,
+    "n_matched": 499
   },
   "dataset": {
-    "gold_images": 120,
-    "pred_images": 120,
+    "gold_images": 203,
+    "pred_images": 203,
     "missing_predictions": 0,
     "iou_threshold": 0.5
   },
-  "gate": {
-    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-2026-09-28/gold_report.json",
-    "f1_delta": -0.0004,
-    "mask_iou_delta": 0.0113,
-    "mask_score": 0.7215,
-    "mask_score_baseline": 0.7103,
-    "matched_only_iou_delta": -0.0033,
-    "class_f1_delta": {
-      "loop_all": 0.0292,
-      "pin": 0.0042
-    },
-    "pass": true
-  }
+  "weights": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-humanonly-2026-09-28/weights.pt",
+  "measured_at": "2026-09-29T17:59:04"
 }
diff --git a/models/registry/yolo26n-seg-alc-v1-mask1-2026-09-28/gold_report.json b/models/registry/yolo26n-seg-alc-v1-mask1-2026-09-28/gold_report.json
index 8fef0b6..2312554 100644
--- a/models/registry/yolo26n-seg-alc-v1-mask1-2026-09-28/gold_report.json
+++ b/models/registry/yolo26n-seg-alc-v1-mask1-2026-09-28/gold_report.json
@@ -1,52 +1,52 @@
 {
   "per_class": {
     "loop_all": {
-      "tp": 112,
-      "fp": 15,
-      "fn": 5,
-      "precision": 0.8818897637795275,
-      "recall": 0.9572649572649573,
-      "f1": 0.9180327868852459,
-      "mean_mask_iou": 0.8879999668937923,
-      "n_matched": 112
+      "tp": 184,
+      "fp": 32,
+      "fn": 8,
+      "precision": 0.8518518518518519,
+      "recall": 0.9583333333333334,
+      "f1": 0.9019607843137256,
+      "mean_mask_iou": 0.8839404070197581,
+      "n_matched": 184
     },
     "pin": {
-      "tp": 84,
-      "fp": 6,
-      "fn": 5,
-      "precision": 0.9333333333333333,
-      "recall": 0.9438202247191011,
-      "f1": 0.9385474860335196,
-      "mean_mask_iou": 0.9283102184111982,
-      "n_matched": 84
+      "tp": 131,
+      "fp": 12,
+      "fn": 15,
+      "precision": 0.916083916083916,
+      "recall": 0.8972602739726028,
+      "f1": 0.9065743944636678,
+      "mean_mask_iou": 0.9297792141522662,
+      "n_matched": 131
     },
     "crystal": {
-      "tp": 16,
-      "fp": 14,
-      "fn": 17,
-      "precision": 0.5333333333333333,
-      "recall": 0.48484848484848486,
-      "f1": 0.507936507936508,
-      "mean_mask_iou": 0.792008917551766,
-      "n_matched": 16
+      "tp": 29,
+      "fp": 27,
+      "fn": 37,
+      "precision": 0.5178571428571429,
+      "recall": 0.4393939393939394,
+      "f1": 0.47540983606557385,
+      "mean_mask_iou": 0.7901618015638238,
+      "n_matched": 29
     },
     "loop_face": {
-      "tp": 85,
-      "fp": 14,
-      "fn": 20,
-      "precision": 0.8585858585858586,
-      "recall": 0.8095238095238095,
-      "f1": 0.8333333333333333,
-      "mean_mask_iou": 0.8686498409776393,
-      "n_matched": 85
+      "tp": 145,
+      "fp": 30,
+      "fn": 27,
+      "precision": 0.8285714285714286,
+      "recall": 0.8430232558139535,
+      "f1": 0.8357348703170029,
+      "mean_mask_iou": 0.8694536352378948,
+      "n_matched": 145
     },
     "ice": {
       "tp": 4,
       "fp": 7,
-      "fn": 28,
+      "fn": 36,
       "precision": 0.36363636363636365,
-      "recall": 0.125,
-      "f1": 0.18604651162790697,
+      "recall": 0.1,
+      "f1": 0.15686274509803924,
       "mean_mask_iou": 0.8357625526706466,
       "n_matched": 4
     },
@@ -62,32 +62,21 @@
     }
   },
   "overall": {
-    "tp": 301,
-    "fp": 56,
-    "fn": 76,
-    "precision": 0.8431372549019608,
-    "recall": 0.7984084880636605,
-    "f1": 0.8201634877384196,
-    "mean_mask_iou": 0.8879883189809155,
-    "n_matched": 301
+    "tp": 493,
+    "fp": 108,
+    "fn": 124,
+    "precision": 0.8202995008319468,
+    "recall": 0.7990275526742301,
+    "f1": 0.8095238095238095,
+    "mean_mask_iou": 0.8859525994140174,
+    "n_matched": 493
   },
   "dataset": {
-    "gold_images": 120,
-    "pred_images": 120,
+    "gold_images": 203,
+    "pred_images": 203,
     "missing_predictions": 0,
     "iou_threshold": 0.5
   },
-  "gate": {
-    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-2026-09-28/gold_report.json",
-    "f1_delta": -0.0003,
-    "mask_iou_delta": -0.0013,
-    "mask_score": 0.709,
-    "mask_score_baseline": 0.7103,
-    "matched_only_iou_delta": 0.0072,
-    "class_f1_delta": {
-      "loop_all": -0.0063,
-      "pin": 0.0196
-    },
-    "pass": false
-  }
+  "weights": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-mask1-2026-09-28/weights.pt",
+  "measured_at": "2026-09-29T17:59:10"
 }
diff --git a/models/registry/yolo26n-seg-alc-v1-noflip-2026-09-28/gold_report.json b/models/registry/yolo26n-seg-alc-v1-noflip-2026-09-28/gold_report.json
index f6355a1..3dc72a6 100644
--- a/models/registry/yolo26n-seg-alc-v1-noflip-2026-09-28/gold_report.json
+++ b/models/registry/yolo26n-seg-alc-v1-noflip-2026-09-28/gold_report.json
@@ -1,52 +1,52 @@
 {
   "per_class": {
     "loop_all": {
-      "tp": 112,
-      "fp": 9,
-      "fn": 5,
-      "precision": 0.9256198347107438,
-      "recall": 0.9572649572649573,
-      "f1": 0.9411764705882353,
-      "mean_mask_iou": 0.8821958664658018,
-      "n_matched": 112
+      "tp": 184,
+      "fp": 18,
+      "fn": 8,
+      "precision": 0.9108910891089109,
+      "recall": 0.9583333333333334,
+      "f1": 0.934010152284264,
+      "mean_mask_iou": 0.8805106649886035,
+      "n_matched": 184
     },
     "pin": {
-      "tp": 84,
-      "fp": 7,
-      "fn": 5,
-      "precision": 0.9230769230769231,
-      "recall": 0.9438202247191011,
-      "f1": 0.9333333333333335,
-      "mean_mask_iou": 0.9220384042537269,
-      "n_matched": 84
+      "tp": 134,
+      "fp": 13,
+      "fn": 12,
+      "precision": 0.9115646258503401,
+      "recall": 0.9178082191780822,
+      "f1": 0.9146757679180887,
+      "mean_mask_iou": 0.9188967633427155,
+      "n_matched": 134
     },
     "crystal": {
-      "tp": 21,
-      "fp": 12,
-      "fn": 12,
-      "precision": 0.6363636363636364,
-      "recall": 0.6363636363636364,
-      "f1": 0.6363636363636364,
-      "mean_mask_iou": 0.8046202903580851,
-      "n_matched": 21
+      "tp": 37,
+      "fp": 23,
+      "fn": 29,
+      "precision": 0.6166666666666667,
+      "recall": 0.5606060606060606,
+      "f1": 0.5873015873015873,
+      "mean_mask_iou": 0.7664666698389817,
+      "n_matched": 37
     },
     "loop_face": {
-      "tp": 88,
-      "fp": 19,
-      "fn": 17,
-      "precision": 0.822429906542056,
-      "recall": 0.8380952380952381,
-      "f1": 0.830188679245283,
-      "mean_mask_iou": 0.8620342125692534,
-      "n_matched": 88
+      "tp": 148,
+      "fp": 36,
+      "fn": 24,
+      "precision": 0.8043478260869565,
+      "recall": 0.8604651162790697,
+      "f1": 0.8314606741573034,
+      "mean_mask_iou": 0.8654994618769852,
+      "n_matched": 148
     },
     "ice": {
       "tp": 3,
       "fp": 6,
-      "fn": 29,
+      "fn": 37,
       "precision": 0.3333333333333333,
-      "recall": 0.09375,
-      "f1": 0.14634146341463417,
+      "recall": 0.075,
+      "f1": 0.12244897959183673,
       "mean_mask_iou": 0.7530200271235444,
       "n_matched": 3
     },
@@ -62,32 +62,21 @@
     }
   },
   "overall": {
-    "tp": 308,
-    "fp": 53,
-    "fn": 69,
-    "precision": 0.853185595567867,
-    "recall": 0.8169761273209549,
-    "f1": 0.8346883468834688,
-    "mean_mask_iou": 0.8807540905404792,
-    "n_matched": 308
+    "tp": 506,
+    "fp": 96,
+    "fn": 111,
+    "precision": 0.840531561461794,
+    "recall": 0.820097244732577,
+    "f1": 0.830188679245283,
+    "mean_mask_iou": 0.8771904661443354,
+    "n_matched": 506
   },
   "dataset": {
-    "gold_images": 120,
-    "pred_images": 120,
+    "gold_images": 203,
+    "pred_images": 203,
     "missing_predictions": 0,
     "iou_threshold": 0.5
   },
-  "gate": {
-    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-2026-09-28/gold_report.json",
-    "f1_delta": 0.0142,
-    "mask_iou_delta": 0.0093,
-    "mask_score": 0.7196,
-    "mask_score_baseline": 0.7103,
-    "matched_only_iou_delta": -0.0001,
-    "class_f1_delta": {
-      "loop_all": 0.0168,
-      "pin": 0.0144
-    },
-    "pass": true
-  }
+  "weights": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-noflip-2026-09-28/weights.pt",
+  "measured_at": "2026-09-29T17:59:16"
 }
diff --git a/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json b/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json
index b1bb63a..8e959e8 100644
--- a/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json
+++ b/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json
@@ -78,5 +78,5 @@
     "iou_threshold": 0.5
   },
   "weights": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/weights.pt",
-  "measured_at": "2026-09-29T17:06:18"
+  "measured_at": "2026-09-29T17:59:23"
 }
diff --git a/models/registry/yolo26n-seg-overlap-false-2026-04-12/gold_report.json b/models/registry/yolo26n-seg-overlap-false-2026-04-12/gold_report.json
index 9bb0c98..d0fa723 100644
--- a/models/registry/yolo26n-seg-overlap-false-2026-04-12/gold_report.json
+++ b/models/registry/yolo26n-seg-overlap-false-2026-04-12/gold_report.json
@@ -1,54 +1,54 @@
 {
   "per_class": {
     "loop_all": {
-      "tp": 96,
-      "fp": 14,
-      "fn": 21,
-      "precision": 0.8727272727272727,
-      "recall": 0.8205128205128205,
-      "f1": 0.8458149779735681,
-      "mean_mask_iou": 0.8908656322562919,
-      "n_matched": 96
+      "tp": 157,
+      "fp": 20,
+      "fn": 35,
+      "precision": 0.8870056497175142,
+      "recall": 0.8177083333333334,
+      "f1": 0.8509485094850948,
+      "mean_mask_iou": 0.9004335749159704,
+      "n_matched": 157
     },
     "pin": {
-      "tp": 77,
-      "fp": 8,
-      "fn": 12,
-      "precision": 0.9058823529411765,
-      "recall": 0.8651685393258427,
-      "f1": 0.8850574712643678,
-      "mean_mask_iou": 0.9507124467978327,
-      "n_matched": 77
+      "tp": 124,
+      "fp": 22,
+      "fn": 22,
+      "precision": 0.8493150684931506,
+      "recall": 0.8493150684931506,
+      "f1": 0.8493150684931505,
+      "mean_mask_iou": 0.955296756052328,
+      "n_matched": 124
     },
     "crystal": {
-      "tp": 19,
-      "fp": 20,
-      "fn": 14,
-      "precision": 0.48717948717948717,
-      "recall": 0.5757575757575758,
-      "f1": 0.5277777777777778,
-      "mean_mask_iou": 0.8456835987950706,
-      "n_matched": 19
+      "tp": 32,
+      "fp": 44,
+      "fn": 34,
+      "precision": 0.42105263157894735,
+      "recall": 0.48484848484848486,
+      "f1": 0.4507042253521127,
+      "mean_mask_iou": 0.8087035162260534,
+      "n_matched": 32
     },
     "loop_face": {
-      "tp": 79,
-      "fp": 15,
-      "fn": 26,
-      "precision": 0.8404255319148937,
-      "recall": 0.7523809523809524,
-      "f1": 0.7939698492462312,
-      "mean_mask_iou": 0.8938253874073545,
-      "n_matched": 79
+      "tp": 129,
+      "fp": 34,
+      "fn": 43,
+      "precision": 0.7914110429447853,
+      "recall": 0.75,
+      "f1": 0.7701492537313434,
+      "mean_mask_iou": 0.9049739687134764,
+      "n_matched": 129
     },
     "ice": {
-      "tp": 12,
-      "fp": 13,
-      "fn": 20,
-      "precision": 0.48,
-      "recall": 0.375,
-      "f1": 0.42105263157894735,
-      "mean_mask_iou": 0.8346283164485953,
-      "n_matched": 12
+      "tp": 13,
+      "fp": 17,
+      "fn": 27,
+      "precision": 0.43333333333333335,
+      "recall": 0.325,
+      "f1": 0.37142857142857144,
+      "mean_mask_iou": 0.8470510637223766,
+      "n_matched": 13
     },
     "needle": {
       "tp": 0,
@@ -62,21 +62,21 @@
     }
   },
   "overall": {
-    "tp": 283,
-    "fp": 70,
-    "fn": 94,
-    "precision": 0.8016997167138811,
-    "recall": 0.7506631299734748,
-    "f1": 0.7753424657534248,
-    "mean_mask_iou": 0.9025572186562107,
-    "n_matched": 283
+    "tp": 455,
+    "fp": 137,
+    "fn": 162,
+    "precision": 0.768581081081081,
+    "recall": 0.7374392220421394,
+    "f1": 0.7526881720430108,
+    "mean_mask_iou": 0.9086960160966134,
+    "n_matched": 455
   },
   "dataset": {
-    "gold_images": 120,
-    "pred_images": 120,
+    "gold_images": 203,
+    "pred_images": 203,
     "missing_predictions": 0,
     "iou_threshold": 0.5
   },
   "weights": "/home/jungfrau/AareLC_studio/models/best_yolo26n-seg-overlap-false_2026-04-12.pt",
-  "measured_at": "2026-09-28T16:51:44"
+  "measured_at": "2026-09-29T17:59:35"
 }
diff --git a/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/gold_report.json b/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/gold_report.json
index 135d8b8..eb58e28 100644
--- a/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/gold_report.json
+++ b/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/gold_report.json
@@ -77,17 +77,6 @@
     "missing_predictions": 0,
     "iou_threshold": 0.5
   },
-  "gate": {
-    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json",
-    "f1_delta": -0.0123,
-    "mask_iou_delta": -0.0039,
-    "mask_score": 0.7298,
-    "mask_score_baseline": 0.7337,
-    "matched_only_iou_delta": -0.0168,
-    "class_f1_delta": {
-      "loop_all": 0.0089,
-      "pin": 0.0021
-    },
-    "pass": false
-  }
+  "weights": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/weights.pt",
+  "measured_at": "2026-09-29T17:59:29"
 }
diff --git a/runs/clean_gold_comparison.json b/runs/clean_gold_comparison.json
index bebdee5..40e2dde 100644
--- a/runs/clean_gold_comparison.json
+++ b/runs/clean_gold_comparison.json
@@ -4,118 +4,188 @@
     "beamline yolo26n-seg-overlap-false-2026-04-12": {
       "per_class": {
         "loop_all": {
-          "tp": 43,
-          "fp": 14,
-          "fn": 96,
-          "precision": 0.7543859649122807,
-          "recall": 0.30935251798561153,
-          "f1": 0.4387755102040817,
-          "mean_mask_iou": 0.8603647813074378,
-          "n_matched": 43
+          "tp": 104,
+          "fp": 20,
+          "fn": 35,
+          "precision": 0.8387096774193549,
+          "recall": 0.7482014388489209,
+          "f1": 0.790874524714829,
+          "mean_mask_iou": 0.892698616936761,
+          "n_matched": 104
         },
         "pin": {
-          "tp": 48,
-          "fp": 7,
-          "fn": 68,
-          "precision": 0.8727272727272727,
-          "recall": 0.41379310344827586,
-          "f1": 0.5614035087719298,
-          "mean_mask_iou": 0.9478639299321778,
-          "n_matched": 48
+          "tp": 95,
+          "fp": 21,
+          "fn": 21,
+          "precision": 0.8189655172413793,
+          "recall": 0.8189655172413793,
+          "f1": 0.8189655172413793,
+          "mean_mask_iou": 0.955256926145264,
+          "n_matched": 95
         },
         "crystal": {
-          "tp": 2,
-          "fp": 15,
-          "fn": 38,
-          "precision": 0.11764705882352941,
-          "recall": 0.05,
-          "f1": 0.07017543859649122,
-          "mean_mask_iou": 0.8492457223853397,
-          "n_matched": 2
+          "tp": 15,
+          "fp": 39,
+          "fn": 25,
+          "precision": 0.2777777777777778,
+          "recall": 0.375,
+          "f1": 0.3191489361702128,
+          "mean_mask_iou": 0.7672677057932032,
+          "n_matched": 15
         },
         "loop_face": {
-          "tp": 30,
-          "fp": 12,
-          "fn": 90,
-          "precision": 0.7142857142857143,
-          "recall": 0.25,
-          "f1": 0.37037037037037035,
-          "mean_mask_iou": 0.8979746795878049,
-          "n_matched": 30
+          "tp": 80,
+          "fp": 31,
+          "fn": 40,
+          "precision": 0.7207207207207207,
+          "recall": 0.6666666666666666,
+          "f1": 0.6926406926406925,
+          "mean_mask_iou": 0.913358459331145,
+          "n_matched": 80
+        },
+        "ice": {
+          "tp": 2,
+          "fp": 8,
+          "fn": 8,
+          "precision": 0.2,
+          "recall": 0.2,
+          "f1": 0.20000000000000004,
+          "mean_mask_iou": 0.944649552595864,
+          "n_matched": 2
+        }
+      },
+      "overall": {
+        "tp": 296,
+        "fp": 119,
+        "fn": 129,
+        "precision": 0.7132530120481928,
+        "recall": 0.6964705882352941,
+        "f1": 0.7047619047619047,
+        "mean_mask_iou": 0.9123549175128535,
+        "n_matched": 296
+      },
+      "dataset": {
+        "gold_images": 149,
+        "pred_images": 203,
+        "missing_predictions": 0,
+        "iou_threshold": 0.5
+      }
+    },
+    "2026-09-28_alc-v1": {
+      "per_class": {
+        "loop_all": {
+          "tp": 128,
+          "fp": 15,
+          "fn": 11,
+          "precision": 0.8951048951048951,
+          "recall": 0.920863309352518,
+          "f1": 0.9078014184397164,
+          "mean_mask_iou": 0.8788121599178862,
+          "n_matched": 128
+        },
+        "pin": {
+          "tp": 107,
+          "fp": 15,
+          "fn": 9,
+          "precision": 0.8770491803278688,
+          "recall": 0.9224137931034483,
+          "f1": 0.8991596638655462,
+          "mean_mask_iou": 0.9123001433431446,
+          "n_matched": 107
+        },
+        "crystal": {
+          "tp": 20,
+          "fp": 17,
+          "fn": 20,
+          "precision": 0.5405405405405406,
+          "recall": 0.5,
+          "f1": 0.5194805194805195,
+          "mean_mask_iou": 0.7492468138856746,
+          "n_matched": 20
+        },
+        "loop_face": {
+          "tp": 93,
+          "fp": 24,
+          "fn": 27,
+          "precision": 0.7948717948717948,
+          "recall": 0.775,
+          "f1": 0.7848101265822786,
+          "mean_mask_iou": 0.8626208972648556,
+          "n_matched": 93
         },
         "ice": {
           "tp": 1,
-          "fp": 4,
+          "fp": 3,
           "fn": 9,
-          "precision": 0.2,
+          "precision": 0.25,
           "recall": 0.1,
-          "f1": 0.13333333333333333,
-          "mean_mask_iou": 0.8931750741839762,
+          "f1": 0.14285714285714288,
+          "mean_mask_iou": 0.8808139534883721,
           "n_matched": 1
         }
       },
       "overall": {
-        "tp": 124,
-        "fp": 52,
-        "fn": 301,
-        "precision": 0.7045454545454546,
-        "recall": 0.2917647058823529,
-        "f1": 0.41264559068219636,
-        "mean_mask_iou": 0.9034198478996224,
-        "n_matched": 124
+        "tp": 349,
+        "fp": 74,
+        "fn": 76,
+        "precision": 0.8250591016548463,
+        "recall": 0.8211764705882353,
+        "f1": 0.8231132075471698,
+        "mean_mask_iou": 0.8773454598396542,
+        "n_matched": 349
       },
       "dataset": {
         "gold_images": 149,
-        "pred_images": 120,
-        "missing_predictions": 83,
+        "pred_images": 203,
+        "missing_predictions": 0,
         "iou_threshold": 0.5
       }
     },
     "2026-09-28_alc-v1-800px": {
       "per_class": {
         "loop_all": {
-          "tp": 59,
-          "fp": 10,
-          "fn": 80,
-          "precision": 0.855072463768116,
-          "recall": 0.4244604316546763,
-          "f1": 0.5673076923076923,
-          "mean_mask_iou": 0.8815339593847193,
-          "n_matched": 59
+          "tp": 129,
+          "fp": 23,
+          "fn": 10,
+          "precision": 0.8486842105263158,
+          "recall": 0.9280575539568345,
+          "f1": 0.8865979381443299,
+          "mean_mask_iou": 0.8856848104236194,
+          "n_matched": 129
         },
         "pin": {
-          "tp": 56,
-          "fp": 5,
-          "fn": 60,
-          "precision": 0.9180327868852459,
-          "recall": 0.4827586206896552,
-          "f1": 0.632768361581921,
-          "mean_mask_iou": 0.9197473537826396,
-          "n_matched": 56
+          "tp": 103,
+          "fp": 15,
+          "fn": 13,
+          "precision": 0.8728813559322034,
+          "recall": 0.8879310344827587,
+          "f1": 0.8803418803418803,
+          "mean_mask_iou": 0.9172489103741318,
+          "n_matched": 103
         },
         "crystal": {
-          "tp": 4,
-          "fp": 2,
-          "fn": 36,
-          "precision": 0.6666666666666666,
-          "recall": 0.1,
-          "f1": 0.1739130434782609,
-          "mean_mask_iou": 0.8425290211607122,
-          "n_matched": 4
+          "tp": 19,
+          "fp": 24,
+          "fn": 21,
+          "precision": 0.4418604651162791,
+          "recall": 0.475,
+          "f1": 0.4578313253012048,
+          "mean_mask_iou": 0.7921412362927195,
+          "n_matched": 19
         },
         "loop_face": {
-          "tp": 30,
-          "fp": 11,
-          "fn": 90,
-          "precision": 0.7317073170731707,
-          "recall": 0.25,
-          "f1": 0.37267080745341613,
-          "mean_mask_iou": 0.8851212986608521,
-          "n_matched": 30
+          "tp": 86,
+          "fp": 23,
+          "fn": 34,
+          "precision": 0.7889908256880734,
+          "recall": 0.7166666666666667,
+          "f1": 0.7510917030567685,
+          "mean_mask_iou": 0.8753083175517496,
+          "n_matched": 86
         },
         "ice": {
           "tp": 0,
-          "fp": 0,
+          "fp": 1,
           "fn": 10,
           "precision": 0.0,
           "recall": 0.0,
@@ -125,133 +195,133 @@
         }
       },
       "overall": {
-        "tp": 149,
-        "fp": 28,
-        "fn": 276,
-        "precision": 0.8418079096045198,
-        "recall": 0.35058823529411764,
-        "f1": 0.49501661129568103,
-        "mean_mask_iou": 0.8955712111409039,
-        "n_matched": 149
+        "tp": 337,
+        "fp": 86,
+        "fn": 88,
+        "precision": 0.7966903073286052,
+        "recall": 0.7929411764705883,
+        "f1": 0.794811320754717,
+        "mean_mask_iou": 0.8874100211044352,
+        "n_matched": 337
       },
       "dataset": {
         "gold_images": 149,
-        "pred_images": 120,
-        "missing_predictions": 83,
+        "pred_images": 203,
+        "missing_predictions": 0,
         "iou_threshold": 0.5
       }
     },
     "2026-09-28_alc-v1-humanonly": {
       "per_class": {
         "loop_all": {
-          "tp": 60,
-          "fp": 5,
-          "fn": 79,
-          "precision": 0.9230769230769231,
-          "recall": 0.4316546762589928,
-          "f1": 0.5882352941176471,
-          "mean_mask_iou": 0.8627592241834525,
-          "n_matched": 60
+          "tp": 128,
+          "fp": 14,
+          "fn": 11,
+          "precision": 0.9014084507042254,
+          "recall": 0.920863309352518,
+          "f1": 0.911032028469751,
+          "mean_mask_iou": 0.8767094784331082,
+          "n_matched": 128
         },
         "pin": {
-          "tp": 55,
-          "fp": 3,
-          "fn": 61,
-          "precision": 0.9482758620689655,
-          "recall": 0.47413793103448276,
-          "f1": 0.6321839080459769,
-          "mean_mask_iou": 0.9206098193697871,
-          "n_matched": 55
+          "tp": 106,
+          "fp": 8,
+          "fn": 10,
+          "precision": 0.9298245614035088,
+          "recall": 0.9137931034482759,
+          "f1": 0.9217391304347825,
+          "mean_mask_iou": 0.9034180292825147,
+          "n_matched": 106
         },
         "crystal": {
-          "tp": 4,
-          "fp": 9,
-          "fn": 36,
-          "precision": 0.3076923076923077,
-          "recall": 0.1,
-          "f1": 0.15094339622641512,
-          "mean_mask_iou": 0.7422849246805395,
-          "n_matched": 4
+          "tp": 22,
+          "fp": 25,
+          "fn": 18,
+          "precision": 0.46808510638297873,
+          "recall": 0.55,
+          "f1": 0.5057471264367817,
+          "mean_mask_iou": 0.7288413011814885,
+          "n_matched": 22
         },
         "loop_face": {
-          "tp": 40,
-          "fp": 14,
-          "fn": 80,
-          "precision": 0.7407407407407407,
-          "recall": 0.3333333333333333,
-          "f1": 0.4597701149425288,
-          "mean_mask_iou": 0.8552237297527754,
-          "n_matched": 40
+          "tp": 92,
+          "fp": 28,
+          "fn": 28,
+          "precision": 0.7666666666666667,
+          "recall": 0.7666666666666667,
+          "f1": 0.7666666666666667,
+          "mean_mask_iou": 0.8661837018243022,
+          "n_matched": 92
         },
         "ice": {
           "tp": 1,
-          "fp": 0,
+          "fp": 1,
           "fn": 9,
-          "precision": 1.0,
+          "precision": 0.5,
           "recall": 0.1,
-          "f1": 0.18181818181818182,
+          "f1": 0.16666666666666669,
           "mean_mask_iou": 0.8895522388059701,
           "n_matched": 1
         }
       },
       "overall": {
-        "tp": 160,
-        "fp": 31,
-        "fn": 265,
-        "precision": 0.837696335078534,
-        "recall": 0.3764705882352941,
-        "f1": 0.5194805194805195,
-        "mean_mask_iou": 0.8779170915249036,
-        "n_matched": 160
+        "tp": 349,
+        "fp": 76,
+        "fn": 76,
+        "precision": 0.8211764705882353,
+        "recall": 0.8211764705882353,
+        "f1": 0.8211764705882352,
+        "mean_mask_iou": 0.8727624234269883,
+        "n_matched": 349
       },
       "dataset": {
         "gold_images": 149,
-        "pred_images": 120,
-        "missing_predictions": 83,
+        "pred_images": 203,
+        "missing_predictions": 0,
         "iou_threshold": 0.5
       }
     },
     "2026-09-28_alc-v1-mask1": {
       "per_class": {
         "loop_all": {
-          "tp": 59,
-          "fp": 12,
-          "fn": 80,
-          "precision": 0.8309859154929577,
-          "recall": 0.4244604316546763,
-          "f1": 0.5619047619047619,
-          "mean_mask_iou": 0.8836420109433821,
-          "n_matched": 59
+          "tp": 131,
+          "fp": 29,
+          "fn": 8,
+          "precision": 0.81875,
+          "recall": 0.9424460431654677,
+          "f1": 0.8762541806020067,
+          "mean_mask_iou": 0.8803352461464905,
+          "n_matched": 131
         },
         "pin": {
-          "tp": 55,
-          "fp": 2,
-          "fn": 61,
-          "precision": 0.9649122807017544,
-          "recall": 0.47413793103448276,
-          "f1": 0.6358381502890174,
-          "mean_mask_iou": 0.9211536151590175,
-          "n_matched": 55
+          "tp": 102,
+          "fp": 8,
+          "fn": 14,
+          "precision": 0.9272727272727272,
+          "recall": 0.8793103448275862,
+          "f1": 0.902654867256637,
+          "mean_mask_iou": 0.9263379170701196,
+          "n_matched": 102
         },
         "crystal": {
-          "tp": 3,
-          "fp": 3,
-          "fn": 37,
+          "tp": 16,
+          "fp": 16,
+          "fn": 24,
           "precision": 0.5,
-          "recall": 0.075,
-          "f1": 0.13043478260869565,
-          "mean_mask_iou": 0.8327909254868483,
-          "n_matched": 3
+          "recall": 0.4,
+          "f1": 0.4444444444444445,
+          "mean_mask_iou": 0.7963076463114489,
+          "n_matched": 16
         },
         "loop_face": {
-          "tp": 39,
-          "fp": 7,
-          "fn": 81,
-          "precision": 0.8478260869565217,
-          "recall": 0.325,
-          "f1": 0.46987951807228917,
-          "mean_mask_iou": 0.8596259543418816,
-          "n_matched": 39
+          "tp": 99,
+          "fp": 23,
+          "fn": 21,
+          "precision": 0.8114754098360656,
+          "recall": 0.825,
+          "f1": 0.8181818181818182,
+          "mean_mask_iou": 0.8662722509669575,
+          "n_matched": 99
         },
         "ice": {
           "tp": 1,
@@ -265,63 +335,63 @@
         }
       },
       "overall": {
-        "tp": 157,
-        "fp": 24,
-        "fn": 268,
-        "precision": 0.8674033149171271,
-        "recall": 0.36941176470588233,
-        "f1": 0.5181518151815181,
-        "mean_mask_iou": 0.8899232217103997,
-        "n_matched": 157
+        "tp": 349,
+        "fp": 76,
+        "fn": 76,
+        "precision": 0.8211764705882353,
+        "recall": 0.8211764705882353,
+        "f1": 0.8211764705882352,
+        "mean_mask_iou": 0.8859830753764693,
+        "n_matched": 349
       },
       "dataset": {
         "gold_images": 149,
-        "pred_images": 120,
-        "missing_predictions": 83,
+        "pred_images": 203,
+        "missing_predictions": 0,
         "iou_threshold": 0.5
       }
     },
     "2026-09-28_alc-v1-noflip": {
       "per_class": {
         "loop_all": {
-          "tp": 59,
-          "fp": 8,
-          "fn": 80,
-          "precision": 0.8805970149253731,
-          "recall": 0.4244604316546763,
-          "f1": 0.5728155339805825,
-          "mean_mask_iou": 0.8707171878261458,
-          "n_matched": 59
+          "tp": 131,
+          "fp": 17,
+          "fn": 8,
+          "precision": 0.8851351351351351,
+          "recall": 0.9424460431654677,
+          "f1": 0.9128919860627178,
+          "mean_mask_iou": 0.8746590793547774,
+          "n_matched": 131
         },
         "pin": {
-          "tp": 55,
-          "fp": 4,
-          "fn": 61,
-          "precision": 0.9322033898305084,
-          "recall": 0.47413793103448276,
-          "f1": 0.6285714285714284,
-          "mean_mask_iou": 0.9138805674783179,
-          "n_matched": 55
+          "tp": 105,
+          "fp": 10,
+          "fn": 11,
+          "precision": 0.9130434782608695,
+          "recall": 0.9051724137931034,
+          "f1": 0.9090909090909092,
+          "mean_mask_iou": 0.9137559194468411,
+          "n_matched": 105
         },
         "crystal": {
-          "tp": 4,
-          "fp": 6,
-          "fn": 36,
-          "precision": 0.4,
-          "recall": 0.1,
-          "f1": 0.16000000000000003,
-          "mean_mask_iou": 0.7713352240501519,
-          "n_matched": 4
+          "tp": 20,
+          "fp": 17,
+          "fn": 20,
+          "precision": 0.5405405405405406,
+          "recall": 0.5,
+          "f1": 0.5194805194805195,
+          "mean_mask_iou": 0.7273790791361571,
+          "n_matched": 20
         },
         "loop_face": {
-          "tp": 40,
-          "fp": 14,
-          "fn": 80,
-          "precision": 0.7407407407407407,
-          "recall": 0.3333333333333333,
-          "f1": 0.4597701149425288,
-          "mean_mask_iou": 0.8528345417122416,
-          "n_matched": 40
+          "tp": 100,
+          "fp": 31,
+          "fn": 20,
+          "precision": 0.7633587786259542,
+          "recall": 0.8333333333333334,
+          "f1": 0.796812749003984,
+          "mean_mask_iou": 0.8634829132018921,
+          "n_matched": 100
         },
         "ice": {
           "tp": 1,
@@ -335,63 +405,63 @@
         }
       },
       "overall": {
-        "tp": 159,
-        "fp": 32,
-        "fn": 266,
-        "precision": 0.8324607329842932,
-        "recall": 0.37411764705882355,
-        "f1": 0.5162337662337663,
-        "mean_mask_iou": 0.8789316384172502,
-        "n_matched": 159
+        "tp": 357,
+        "fp": 75,
+        "fn": 68,
+        "precision": 0.8263888888888888,
+        "recall": 0.84,
+        "f1": 0.8331388564760793,
+        "mean_mask_iou": 0.8748914467532463,
+        "n_matched": 357
       },
       "dataset": {
         "gold_images": 149,
-        "pred_images": 120,
-        "missing_predictions": 83,
+        "pred_images": 203,
+        "missing_predictions": 0,
         "iou_threshold": 0.5
       }
     },
-    "2026-09-28_alc-v1": {
+    "2026-09-29_alc-v1-noflip-mask1": {
       "per_class": {
         "loop_all": {
-          "tp": 58,
-          "fp": 7,
-          "fn": 81,
-          "precision": 0.8923076923076924,
-          "recall": 0.4172661870503597,
-          "f1": 0.5686274509803922,
-          "mean_mask_iou": 0.873542286431727,
-          "n_matched": 58
+          "tp": 130,
+          "fp": 20,
+          "fn": 9,
+          "precision": 0.8666666666666667,
+          "recall": 0.935251798561151,
+          "f1": 0.8996539792387542,
+          "mean_mask_iou": 0.8844486511300187,
+          "n_matched": 130
         },
         "pin": {
-          "tp": 56,
-          "fp": 7,
-          "fn": 60,
-          "precision": 0.8888888888888888,
-          "recall": 0.4827586206896552,
-          "f1": 0.6256983240223464,
-          "mean_mask_iou": 0.913231968211653,
-          "n_matched": 56
+          "tp": 105,
+          "fp": 10,
+          "fn": 11,
+          "precision": 0.9130434782608695,
+          "recall": 0.9051724137931034,
+          "f1": 0.9090909090909092,
+          "mean_mask_iou": 0.9320705666956062,
+          "n_matched": 105
         },
         "crystal": {
-          "tp": 4,
-          "fp": 2,
-          "fn": 36,
-          "precision": 0.6666666666666666,
-          "recall": 0.1,
-          "f1": 0.1739130434782609,
-          "mean_mask_iou": 0.7662505101273234,
-          "n_matched": 4
+          "tp": 21,
+          "fp": 31,
+          "fn": 19,
+          "precision": 0.40384615384615385,
+          "recall": 0.525,
+          "f1": 0.4565217391304348,
+          "mean_mask_iou": 0.7718891071431425,
+          "n_matched": 21
         },
         "loop_face": {
-          "tp": 38,
-          "fp": 9,
-          "fn": 82,
-          "precision": 0.8085106382978723,
-          "recall": 0.31666666666666665,
-          "f1": 0.45508982035928147,
-          "mean_mask_iou": 0.8637306298851367,
-          "n_matched": 38
+          "tp": 98,
+          "fp": 33,
+          "fn": 22,
+          "precision": 0.7480916030534351,
+          "recall": 0.8166666666666667,
+          "f1": 0.7808764940239045,
+          "mean_mask_iou": 0.875341660411756,
+          "n_matched": 98
         },
         "ice": {
           "tp": 1,
@@ -400,94 +470,24 @@
           "precision": 0.5,
           "recall": 0.1,
           "f1": 0.16666666666666669,
-          "mean_mask_iou": 0.8808139534883721,
-          "n_matched": 1
-        }
-      },
-      "overall": {
-        "tp": 157,
-        "fp": 26,
-        "fn": 268,
-        "precision": 0.8579234972677595,
-        "recall": 0.36941176470588233,
-        "f1": 0.5164473684210525,
-        "mean_mask_iou": 0.8826370876593987,
-        "n_matched": 157
-      },
-      "dataset": {
-        "gold_images": 149,
-        "pred_images": 120,
-        "missing_predictions": 83,
-        "iou_threshold": 0.5
-      }
-    },
-    "2026-09-29_alc-v1-noflip-mask1": {
-      "per_class": {
-        "loop_all": {
-          "tp": 59,
-          "fp": 7,
-          "fn": 80,
-          "precision": 0.8939393939393939,
-          "recall": 0.4244604316546763,
-          "f1": 0.5756097560975609,
-          "mean_mask_iou": 0.8814978671212703,
-          "n_matched": 59
-        },
-        "pin": {
-          "tp": 55,
-          "fp": 4,
-          "fn": 61,
-          "precision": 0.9322033898305084,
-          "recall": 0.47413793103448276,
-          "f1": 0.6285714285714284,
-          "mean_mask_iou": 0.9269541978233039,
-          "n_matched": 55
-        },
-        "crystal": {
-          "tp": 4,
-          "fp": 4,
-          "fn": 36,
-          "precision": 0.5,
-          "recall": 0.1,
-          "f1": 0.16666666666666669,
-          "mean_mask_iou": 0.794146590637362,
-          "n_matched": 4
-        },
-        "loop_face": {
-          "tp": 38,
-          "fp": 12,
-          "fn": 82,
-          "precision": 0.76,
-          "recall": 0.31666666666666665,
-          "f1": 0.4470588235294118,
-          "mean_mask_iou": 0.872694452002524,
-          "n_matched": 38
-        },
-        "ice": {
-          "tp": 1,
-          "fp": 0,
-          "fn": 9,
-          "precision": 1.0,
-          "recall": 0.1,
-          "f1": 0.18181818181818182,
           "mean_mask_iou": 0.901840490797546,
           "n_matched": 1
         }
       },
       "overall": {
-        "tp": 157,
-        "fp": 27,
-        "fn": 268,
-        "precision": 0.8532608695652174,
-        "recall": 0.36941176470588233,
-        "f1": 0.51559934318555,
-        "mean_mask_iou": 0.8931953571329908,
-        "n_matched": 157
+        "tp": 355,
+        "fp": 95,
+        "fn": 70,
+        "precision": 0.7888888888888889,
+        "recall": 0.8352941176470589,
+        "f1": 0.8114285714285714,
+        "mean_mask_iou": 0.88941050312985,
+        "n_matched": 355
       },
       "dataset": {
         "gold_images": 149,
-        "pred_images": 120,
-        "missing_predictions": 83,
+        "pred_images": 203,
+        "missing_predictions": 0,
         "iou_threshold": 0.5
       }
     },
diff --git a/runs/clean_gold_comparison.txt b/runs/clean_gold_comparison.txt
index 46a48d4..801224c 100644
--- a/runs/clean_gold_comparison.txt
+++ b/runs/clean_gold_comparison.txt
@@ -1,8 +1,10 @@
-clean gold frames: 66
+clean gold frames: 149
 run                                                F1  recall  IoU matched  mask score   loop/pin/crystal/face/ice F1
-beamline yolo26n-seg-overlap-false-2026-04-12   0.687    0.67        0.903       0.606   0.71/0.84/0.17/0.63/0.29
-2026-09-28_alc-v1-800px                         0.823    0.81        0.896       0.721   0.89/0.93/0.62/0.64/0.00
-2026-09-28_alc-v1-humanonly                     0.851    0.86        0.878       0.759   0.93/0.94/0.40/0.75/0.67
-2026-09-28_alc-v1-mask1                         0.858    0.85        0.890       0.755   0.87/0.95/0.46/0.79/0.67
-2026-09-28_alc-v1-noflip                        0.846    0.86        0.879       0.755   0.90/0.93/0.47/0.75/0.67
-2026-09-28_alc-v1                               0.853    0.85        0.883       0.749   0.90/0.92/0.62/0.76/0.50
+beamline yolo26n-seg-overlap-false-2026-04-12   0.705    0.70        0.912       0.635   0.79/0.82/0.32/0.69/0.20
+2026-09-28_alc-v1                               0.823    0.82        0.877       0.720   0.91/0.90/0.52/0.78/0.14
+2026-09-28_alc-v1-800px                         0.795    0.79        0.887       0.704   0.89/0.88/0.46/0.75/0.00
+2026-09-28_alc-v1-humanonly                     0.821    0.82        0.873       0.717   0.91/0.92/0.51/0.77/0.17
+2026-09-28_alc-v1-mask1                         0.821    0.82        0.886       0.728   0.88/0.90/0.44/0.82/0.18
+2026-09-28_alc-v1-noflip                        0.833    0.84        0.875       0.735   0.91/0.91/0.52/0.80/0.18
+2026-09-29_alc-v1-noflip-mask1                  0.811    0.84        0.889       0.743   0.90/0.91/0.46/0.78/0.17
+2026-09-29_v2-recipeA                           0.814    0.84        0.877       0.741   0.91/0.90/0.47/0.77/0.12