diff --git a/media/RESULTS.md b/media/RESULTS.md
index b275b88..6cd94d0 100644
--- a/media/RESULTS.md
+++ b/media/RESULTS.md
@@ -294,91 +294,91 @@ Results of the model in various experiments on different datasets. \
 
 ```json
 {
-  "img_loading": 0.042184,
-  "demosaicing": 0.000678806,
-  "avg_time_2d": 0.0147901,
-  "avg_time_3d": 0.000119385,
-  "fps": 64.1505
+  "img_loading": 0.0422866,
+  "demosaicing": 0.000699368,
+  "avg_time_2d": 0.014883,
+  "avg_time_3d": 0.00011746,
+  "fps": 63.6951
 }
 {
   "triangulator_calls": 301,
-  "init_time": 2.93031e-06,
-  "undistort_time": 1.43119e-05,
-  "project_time": 2.3093e-06,
-  "match_time": 8.56559e-06,
-  "pairs_time": 4.58829e-06,
-  "pair_scoring_time": 2.63603e-05,
-  "grouping_time": 4.53817e-06,
-  "full_time": 2.74936e-05,
-  "merge_time": 1.15484e-05,
-  "post_time": 6.9437e-06,
-  "convert_time": 1.26757e-07,
-  "total_time": 0.000109953
+  "init_time": 2.91101e-06,
+  "undistort_time": 1.43844e-05,
+  "project_time": 2.16363e-06,
+  "match_time": 8.39535e-06,
+  "pairs_time": 4.6119e-06,
+  "pair_scoring_time": 2.70302e-05,
+  "grouping_time": 4.49531e-06,
+  "full_time": 2.73468e-05,
+  "merge_time": 1.2003e-05,
+  "post_time": 5.09957e-06,
+  "convert_time": 1.25462e-07,
+  "total_time": 0.000108806
 }
 {
   "person_nums": {
     "total_frames": 301,
     "total_labels": 477,
-    "total_preds": 832,
+    "total_preds": 827,
     "considered_empty": 0,
     "valid_preds": 477,
-    "invalid_preds": 355,
+    "invalid_preds": 350,
     "missing": 0,
-    "invalid_fraction": 0.42668,
-    "precision": 0.57332,
+    "invalid_fraction": 0.42322,
+    "precision": 0.57678,
     "recall": 1.0,
-    "f1": 0.7288,
-    "non_empty": 832
+    "f1": 0.7316,
+    "non_empty": 827
   },
   "mpjpe": {
     "count": 477,
-    "mean": 0.047771,
-    "median": 0.04252,
-    "std": 0.014668,
-    "sem": 0.000672,
-    "min": 0.029325,
-    "max": 0.105412,
+    "mean": 0.047577,
+    "median": 0.042522,
+    "std": 0.014987,
+    "sem": 0.000687,
+    "min": 0.028778,
+    "max": 0.118939,
     "recall-0.025": 0.0,
-    "recall-0.05": 0.702306,
+    "recall-0.05": 0.72327,
     "recall-0.1": 0.987421,
     "recall-0.15": 1.0,
     "recall-0.25": 1.0,
     "recall-0.5": 1.0,
     "num_labels": 477,
     "ap-0.025": 0.0,
-    "ap-0.05": 0.429122,
-    "ap-0.1": 0.750799,
-    "ap-0.15": 0.760584,
-    "ap-0.25": 0.760584,
-    "ap-0.5": 0.760584
+    "ap-0.05": 0.45527,
+    "ap-0.1": 0.751931,
+    "ap-0.15": 0.761005,
+    "ap-0.25": 0.761005,
+    "ap-0.5": 0.761005
   },
   "head": {
     "count": 477,
-    "mean": 0.053874,
-    "median": 0.04981,
-    "std": 0.024502,
-    "sem": 0.001123,
-    "min": 0.005023,
-    "max": 0.171118,
-    "recall-0.025": 0.090147,
-    "recall-0.05": 0.505241,
-    "recall-0.1": 0.939203,
-    "recall-0.15": 0.997904,
+    "mean": 0.052991,
+    "median": 0.048986,
+    "std": 0.025142,
+    "sem": 0.001152,
+    "min": 0.006559,
+    "max": 0.183673,
+    "recall-0.025": 0.106918,
+    "recall-0.05": 0.522013,
+    "recall-0.1": 0.937107,
+    "recall-0.15": 0.993711,
     "recall-0.25": 1.0,
     "recall-0.5": 1.0,
     "num_labels": 477
   },
   "shoulder_left": {
     "count": 477,
-    "mean": 0.042207,
-    "median": 0.036661,
-    "std": 0.020587,
-    "sem": 0.000944,
+    "mean": 0.041956,
+    "median": 0.037041,
+    "std": 0.02074,
+    "sem": 0.000951,
     "min": 0.004368,
-    "max": 0.136632,
-    "recall-0.025": 0.165618,
-    "recall-0.05": 0.740042,
-    "recall-0.1": 0.983229,
+    "max": 0.132918,
+    "recall-0.025": 0.174004,
+    "recall-0.05": 0.754717,
+    "recall-0.1": 0.985325,
     "recall-0.15": 1.0,
     "recall-0.25": 1.0,
     "recall-0.5": 1.0,
@@ -386,14 +386,14 @@ Results of the model in various experiments on different datasets. \
   },
   "shoulder_right": {
     "count": 477,
-    "mean": 0.049707,
-    "median": 0.046776,
-    "std": 0.022703,
-    "sem": 0.001041,
+    "mean": 0.049604,
+    "median": 0.046469,
+    "std": 0.022881,
+    "sem": 0.001049,
     "min": 0.005334,
-    "max": 0.149673,
-    "recall-0.025": 0.096436,
-    "recall-0.05": 0.551363,
+    "max": 0.148157,
+    "recall-0.025": 0.100629,
+    "recall-0.05": 0.561845,
     "recall-0.1": 0.955975,
     "recall-0.15": 1.0,
     "recall-0.25": 1.0,
@@ -402,15 +402,15 @@ Results of the model in various experiments on different datasets. \
   },
   "elbow_left": {
     "count": 477,
-    "mean": 0.040814,
-    "median": 0.032286,
-    "std": 0.029281,
-    "sem": 0.001342,
-    "min": 0.003459,
+    "mean": 0.040574,
+    "median": 0.03197,
+    "std": 0.028864,
+    "sem": 0.001323,
+    "min": 0.003618,
     "max": 0.309844,
-    "recall-0.025": 0.318658,
-    "recall-0.05": 0.748428,
-    "recall-0.1": 0.951782,
+    "recall-0.025": 0.312369,
+    "recall-0.05": 0.754717,
+    "recall-0.1": 0.953878,
     "recall-0.15": 0.997904,
     "recall-0.25": 0.997904,
     "recall-0.5": 1.0,
@@ -418,31 +418,31 @@ Results of the model in various experiments on different datasets. \
   },
   "elbow_right": {
     "count": 477,
-    "mean": 0.053474,
-    "median": 0.045174,
-    "std": 0.040287,
-    "sem": 0.001847,
-    "min": 0.00327,
-    "max": 0.244516,
-    "recall-0.025": 0.253669,
-    "recall-0.05": 0.553459,
-    "recall-0.1": 0.903564,
-    "recall-0.15": 0.960168,
+    "mean": 0.053198,
+    "median": 0.043315,
+    "std": 0.040536,
+    "sem": 0.001858,
+    "min": 0.003567,
+    "max": 0.242994,
+    "recall-0.025": 0.27044,
+    "recall-0.05": 0.563941,
+    "recall-0.1": 0.897275,
+    "recall-0.15": 0.962264,
     "recall-0.25": 1.0,
     "recall-0.5": 1.0,
     "num_labels": 477
   },
   "wrist_left": {
     "count": 477,
-    "mean": 0.060043,
-    "median": 0.054688,
-    "std": 0.038061,
-    "sem": 0.001745,
+    "mean": 0.059301,
+    "median": 0.054034,
+    "std": 0.038637,
+    "sem": 0.001771,
     "min": 0.00164,
-    "max": 0.323995,
-    "recall-0.025": 0.134172,
-    "recall-0.05": 0.396226,
-    "recall-0.1": 0.907757,
+    "max": 0.32449,
+    "recall-0.025": 0.146751,
+    "recall-0.05": 0.412998,
+    "recall-0.1": 0.909853,
     "recall-0.15": 0.964361,
     "recall-0.25": 0.991614,
     "recall-0.5": 1.0,
@@ -450,47 +450,47 @@ Results of the model in various experiments on different datasets. \
   },
   "wrist_right": {
     "count": 477,
-    "mean": 0.058016,
-    "median": 0.054634,
-    "std": 0.030706,
-    "sem": 0.001407,
+    "mean": 0.058286,
+    "median": 0.053472,
+    "std": 0.034178,
+    "sem": 0.001567,
     "min": 0.008622,
-    "max": 0.238817,
-    "recall-0.025": 0.111111,
-    "recall-0.05": 0.421384,
-    "recall-0.1": 0.903564,
-    "recall-0.15": 0.985325,
-    "recall-0.25": 1.0,
+    "max": 0.382984,
+    "recall-0.025": 0.113208,
+    "recall-0.05": 0.440252,
+    "recall-0.1": 0.90566,
+    "recall-0.15": 0.981132,
+    "recall-0.25": 0.997904,
     "recall-0.5": 1.0,
     "num_labels": 477
   },
   "hip_left": {
     "count": 477,
-    "mean": 0.048043,
+    "mean": 0.047666,
     "median": 0.042435,
-    "std": 0.025895,
-    "sem": 0.001187,
-    "min": 0.006314,
-    "max": 0.173228,
-    "recall-0.025": 0.159329,
-    "recall-0.05": 0.624738,
-    "recall-0.1": 0.955975,
-    "recall-0.15": 0.997904,
+    "std": 0.025966,
+    "sem": 0.00119,
+    "min": 0.008125,
+    "max": 0.145941,
+    "recall-0.025": 0.188679,
+    "recall-0.05": 0.628931,
+    "recall-0.1": 0.958071,
+    "recall-0.15": 1.0,
     "recall-0.25": 1.0,
     "recall-0.5": 1.0,
     "num_labels": 477
   },
   "hip_right": {
     "count": 477,
-    "mean": 0.057878,
-    "median": 0.056579,
-    "std": 0.023899,
+    "mean": 0.057617,
+    "median": 0.058504,
+    "std": 0.023884,
     "sem": 0.001095,
     "min": 0.006301,
     "max": 0.132422,
-    "recall-0.025": 0.098532,
-    "recall-0.05": 0.352201,
-    "recall-0.1": 0.9413,
+    "recall-0.025": 0.102725,
+    "recall-0.05": 0.354298,
+    "recall-0.1": 0.943396,
     "recall-0.15": 1.0,
     "recall-0.25": 1.0,
     "recall-0.5": 1.0,
@@ -498,15 +498,15 @@ Results of the model in various experiments on different datasets. \
   },
   "knee_left": {
     "count": 477,
-    "mean": 0.040068,
-    "median": 0.036879,
-    "std": 0.02442,
-    "sem": 0.001119,
-    "min": 0.005242,
+    "mean": 0.040292,
+    "median": 0.037394,
+    "std": 0.024591,
+    "sem": 0.001127,
+    "min": 0.00468,
     "max": 0.190112,
-    "recall-0.025": 0.257862,
-    "recall-0.05": 0.754717,
-    "recall-0.1": 0.974843,
+    "recall-0.025": 0.264151,
+    "recall-0.05": 0.752621,
+    "recall-0.1": 0.972746,
     "recall-0.15": 0.989518,
     "recall-0.25": 1.0,
     "recall-0.5": 1.0,
@@ -514,14 +514,14 @@ Results of the model in various experiments on different datasets. \
   },
   "knee_right": {
     "count": 477,
-    "mean": 0.039748,
-    "median": 0.035762,
-    "std": 0.023,
-    "sem": 0.001054,
-    "min": 0.007276,
+    "mean": 0.039685,
+    "median": 0.035562,
+    "std": 0.023057,
+    "sem": 0.001057,
+    "min": 0.005833,
     "max": 0.184978,
-    "recall-0.025": 0.316562,
-    "recall-0.05": 0.72956,
+    "recall-0.025": 0.318658,
+    "recall-0.05": 0.735849,
     "recall-0.1": 0.974843,
     "recall-0.15": 0.997904,
     "recall-0.25": 1.0,
@@ -530,13 +530,13 @@ Results of the model in various experiments on different datasets. \
   },
   "ankle_left": {
     "count": 477,
-    "mean": 0.036387,
-    "median": 0.028113,
-    "std": 0.030604,
-    "sem": 0.001403,
+    "mean": 0.036424,
+    "median": 0.027617,
+    "std": 0.030554,
+    "sem": 0.0014,
     "min": 0.004893,
     "max": 0.223793,
-    "recall-0.025": 0.429769,
+    "recall-0.025": 0.42348,
     "recall-0.05": 0.819706,
     "recall-0.1": 0.947589,
     "recall-0.15": 0.983229,
@@ -546,14 +546,14 @@ Results of the model in various experiments on different datasets. \
   },
   "ankle_right": {
     "count": 477,
-    "mean": 0.040761,
-    "median": 0.030879,
-    "std": 0.037437,
-    "sem": 0.001716,
-    "min": 0.003316,
+    "mean": 0.040909,
+    "median": 0.030845,
+    "std": 0.037585,
+    "sem": 0.001723,
+    "min": 0.003445,
     "max": 0.28852,
     "recall-0.025": 0.301887,
-    "recall-0.05": 0.809224,
+    "recall-0.05": 0.807128,
     "recall-0.1": 0.930818,
     "recall-0.15": 0.968553,
     "recall-0.25": 0.997904,
@@ -562,18 +562,18 @@ Results of the model in various experiments on different datasets. \
   },
   "joint_recalls": {
     "num_labels": 6201,
-    "recall-0.025": 0.20916,
-    "recall-0.05": 0.6149,
-    "recall-0.1": 0.94356,
-    "recall-0.15": 0.98774,
-    "recall-0.25": 0.99887,
+    "recall-0.025": 0.21609,
+    "recall-0.05": 0.62329,
+    "recall-0.1": 0.94291,
+    "recall-0.15": 0.9871,
+    "recall-0.25": 0.99871,
     "recall-0.5": 1.0
   }
 }
 {
   "total_parts": 6678,
-  "correct_parts": 6621,
-  "pcp": 0.991465
+  "correct_parts": 6623,
+  "pcp": 0.991764
 }
 ```
 
diff --git a/rpt/triangulator.cpp b/rpt/triangulator.cpp
index 098e6ed..a538e41 100644
--- a/rpt/triangulator.cpp
+++ b/rpt/triangulator.cpp
@@ -525,6 +525,7 @@ std::vector<std::vector<std::array<float, 4>>> TriangulatorInternal::triangulate
     }
 
     // Convert inputs to internal formats
+    this->num_cams = cameras.size();
     std::vector<std::vector<std::vector<std::array<float, 3>>>> i_poses_2d = poses_2d;
     std::vector<CameraInternal> internal_cameras;
     for (size_t i = 0; i < cameras.size(); ++i)
@@ -1809,6 +1810,7 @@ std::vector<std::array<float, 4>> TriangulatorInternal::merge_group(
     // Use only those triangulations with a high score
     std::vector<std::array<float, 4>> sum_poses(num_joints, {0.0, 0.0, 0.0, 0.0});
     std::vector<float> sum_mask1(num_joints, 0);
+    float offset1 = (1.0 - min_score) / 2.0;
     for (size_t i = 0; i < num_poses; ++i)
     {
         const auto &pose = poses_3d[i];
@@ -1816,7 +1818,7 @@ std::vector<std::array<float, 4>> TriangulatorInternal::merge_group(
         for (size_t j = 0; j < num_joints; ++j)
         {
             float score = pose[j][3];
-            if (score > min_score)
+            if (score > min_score - offset1)
             {
                 float weight = std::pow(score, 2);
                 sum_poses[j][0] += pose[j][0] * weight;
@@ -1847,7 +1849,7 @@ std::vector<std::array<float, 4>> TriangulatorInternal::merge_group(
     for (size_t j = 0; j < num_joints; ++j)
     {
         float score = initial_pose_3d[j][3];
-        if (score > min_score)
+        if (score > min_score - offset1)
         {
             jmask[j] = true;
             center[0] += initial_pose_3d[j][0];
@@ -1875,7 +1877,7 @@ std::vector<std::array<float, 4>> TriangulatorInternal::merge_group(
     }
 
     // Drop joints with low scores and filter outlying joints using distance threshold
-    float offset = 0.1;
+    float offset2 = (1.0 - min_score) * 2.0;
     float max_dist_sq = 1.2 * 1.2;
     std::vector<std::vector<bool>> mask(num_poses, std::vector<bool>(num_joints, false));
     std::vector<std::vector<float>> distances(num_poses, std::vector<float>(num_joints, 0.0));
@@ -1891,7 +1893,7 @@ std::vector<std::array<float, 4>> TriangulatorInternal::merge_group(
             distances[i][j] = dist_sq;
 
             float score = pose[j][3];
-            if (dist_sq <= max_dist_sq && score > (min_score - offset))
+            if (dist_sq <= max_dist_sq && score > (min_score - offset2))
             {
                 mask[i][j] = true;
             }
@@ -1899,7 +1901,7 @@ std::vector<std::array<float, 4>> TriangulatorInternal::merge_group(
     }
 
     // Select the best-k proposals for each joint that are closest to the initial pose
-    int keep_best = std::max(3, (int)num_poses / 3);
+    int keep_best = std::max((int)std::ceil(this->num_cams / 2.0), (int)std::ceil(num_poses / 3.0));
     std::vector<std::vector<bool>> best_k_mask(num_poses, std::vector<bool>(num_joints, false));
     for (size_t j = 0; j < num_joints; ++j)
     {
diff --git a/rpt/triangulator.hpp b/rpt/triangulator.hpp
index 62c3615..b87aefa 100644
--- a/rpt/triangulator.hpp
+++ b/rpt/triangulator.hpp
@@ -54,6 +54,7 @@ public:
 private:
     float min_match_score;
     float min_group_size;
+    size_t num_cams;
 
     const std::vector<std::string> core_joints = {
         "shoulder_left",