Some fixes. Updated benchmarks.

scripts/test_skelda_dataset.py

@@ -384,7 +384,7 @@ def main():
             poses3D = np.zeros([1, len(joint_names_3d), 4]).tolist()
         else:
             spt_cameras = spt.convert_cameras(label["cameras"])
-            roomparams = [label["room_center"], label["room_size"]]
+            roomparams = [label["room_size"], label["room_center"]]
             poses3D = triangulator.triangulate_poses(
                 poses_2d, spt_cameras, roomparams, joint_names_2d
             )

spt/interface.hpp

@@ -30,7 +30,7 @@ public:
      *
      * @param poses_2d List of shape [views, persons, joints, 3], containing the 2D poses.
      * @param cameras List of cameras.
-     * @param roomparams Room parameters (room center, room size).
+     * @param roomparams Room parameters (room size, room center).
      * @param joint_names List of 2D joint names.
      *
      * @return List of shape [persons, joints, 4], containing the 3D poses.

spt/triangulator.cpp

@@ -925,8 +925,8 @@ std::pair<cv::Mat, float> TriangulatorInternal::triangulate_and_score(
     {
         mean[j] *= inv_num_vis;
     }
-    const std::array<float, 3> &center = roomparams[0];
+    const std::array<float, 3> &size = roomparams[0];
-    const std::array<float, 3> &size = roomparams[1];
+    const std::array<float, 3> &center = roomparams[1];
     for (int j = 0; j < 3; ++j)
     {
         if (mean[j] > center[j] + size[j] / 2.0 ||
@@ -1061,9 +1061,7 @@ std::vector<std::tuple<cv::Point3d, cv::Mat, std::vector<int>>> TriangulatorInte
 {
     float max_center_dist_sq = 0.6 * 0.6;
     float max_joint_avg_dist = 0.3;
-
     size_t num_pairs = all_pairs.size();
-    size_t num_joints = all_scored_poses[0].first.rows;
 
     // Calculate pose centers
     std::vector<cv::Point3d> centers;
@@ -1071,6 +1069,8 @@ std::vector<std::tuple<cv::Point3d, cv::Mat, std::vector<int>>> TriangulatorInte
     for (size_t i = 0; i < num_pairs; ++i)
     {
         const cv::Mat &pose_3d = all_scored_poses[i].first;
+        size_t num_joints = pose_3d.rows;
+
         cv::Point3d center(0, 0, 0);
         size_t num_valid = 0;
         for (size_t j = 0; j < num_joints; ++j)
@@ -1102,6 +1102,8 @@ std::vector<std::tuple<cv::Point3d, cv::Mat, std::vector<int>>> TriangulatorInte
     for (size_t i = 0; i < num_pairs; ++i)
     {
         const cv::Mat &pose_3d = all_scored_poses[i].first;
+        size_t num_joints = pose_3d.rows;
+
         const cv::Point3d &center = centers[i];
         float best_dist = std::numeric_limits<float>::infinity();
         int best_group = -1;
@@ -1512,7 +1514,7 @@ void TriangulatorInternal::filter_poses(
         {
             // Check if the mean position is outside the room boundaries
             if (mean[j] > room_half_size[j] + room_center[j] ||
-                mean[j] < -room_half_size[j] - room_center[j])
+                mean[j] < -room_half_size[j] + room_center[j])
             {
                 outside = true;
                 break;
@@ -1522,7 +1524,7 @@ void TriangulatorInternal::filter_poses(
         {
             // Check if any limb is too far outside the room
             if (maxs[j] > room_half_size[j] + room_center[j] + wdist ||
-                mins[j] < -room_half_size[j] - room_center[j] - wdist)
+                mins[j] < -room_half_size[j] + room_center[j] - wdist)
             {
                 outside = true;
                 break;
@@ -1540,8 +1542,8 @@ void TriangulatorInternal::filter_poses(
             size_t found_limbs = 0;
             for (size_t j = 0; j < core_limbs_idx.size(); ++j)
             {
-                size_t start_idx = core_limbs_idx[j][0];
+                size_t start_idx = core_joint_idx[core_limbs_idx[j][0]];
-                size_t end_idx = core_limbs_idx[j][1];
+                size_t end_idx = core_joint_idx[core_limbs_idx[j][1]];
 
                 float *joint_start_ptr = pose.ptr<float>(start_idx);
                 float *joint_end_ptr = pose.ptr<float>(end_idx);
@@ -1565,8 +1567,8 @@ void TriangulatorInternal::filter_poses(
         int total_limbs = 0;
         for (size_t j = 0; j < core_limbs_idx.size(); ++j)
         {
-            size_t start_idx = core_limbs_idx[j][0];
+            size_t start_idx = core_joint_idx[core_limbs_idx[j][0]];
-            size_t end_idx = core_limbs_idx[j][1];
+            size_t end_idx = core_joint_idx[core_limbs_idx[j][1]];
 
             float *joint_start_ptr = pose.ptr<float>(start_idx);
             float *joint_end_ptr = pose.ptr<float>(end_idx);
@@ -1727,7 +1729,7 @@ void TriangulatorInternal::add_missing_joints(
                 if (adjacents.find(adname) != adjacents.end())
                 {
                     // Joint is missing; attempt to estimate its position based on adjacent joints
-                    std::array<float, 3> adjacent_position;
+                    std::array<float, 3> adjacent_position = {0.0f, 0.0f, 0.0f};
                     size_t adjacent_count = 0;
 
                     const std::vector<std::string> &adjacent_joint_names = adjacents[adname];

tests/test_interface.py

@@ -26,7 +26,7 @@ def main():
     print("")
 
     # Load input data
-    roomparams = [[0, 0, 1.0], [4.8, 6.0, 2.0]]
+    roomparams = [[4.8, 6.0, 2.0], [0, 0, 1.0]]
     joint_names = [
         "nose",
         "eye_left",
@@ -70,7 +70,7 @@ def main():
     print("")
 
     # Load input data
-    roomparams = [[0, -0.5, 1.2], [5.6, 6.4, 2.4]]
+    roomparams = [[5.6, 6.4, 2.4], [0, -0.5, 1.2]]
     cpath = "/SimplePoseTriangulation/data/p1/sample.json"
     ppath = "/SimplePoseTriangulation/tests/poses_p1.json"
     with open(cpath, "r") as file: