added clustering and merge_th=100

handcrafted_solution.py

@@ -13,6 +13,7 @@ from hoho.color_mappings import gestalt_color_mapping
 from hoho.read_write_colmap import read_cameras_binary, read_images_binary, read_points3D_binary
 from scipy.spatial import KDTree
 from scipy.spatial.distance import cdist
+import scipy.cluster.hierarchy as shc
 
 apex_color = gestalt_color_mapping["apex"]
 eave_end_point = gestalt_color_mapping["eave_end_point"]
@@ -437,6 +438,33 @@ def predict(entry, visualize=False, scale_estimation_coefficient=2.5, **kwargs)
     image_dict = {}
     for k, v in entry["images"].items():
         image_dict[v.name] = v
+    points = [v.xyz for k, v in entry["points3d"].items()]
+    too_big = len(points) > 25000
+
+    if not too_big:
+        points = np.array(points)
+        point_keys = [k for k, v in entry["points3d"].items()]
+        point_keys = np.array(point_keys)
+
+        # print(len(points))
+
+        clustered = shc.fclusterdata(points, 100, criterion='distance')
+        clustered_indices = np.argsort(clustered)
+
+        points = points[clustered_indices]
+        point_keys = point_keys[clustered_indices]
+        clustered = clustered[clustered_indices]
+
+        _, cluster_indices = np.unique(clustered, return_index=True)
+
+        clustered_points = np.split(points, cluster_indices[1:])
+        clustered_keys = np.split(point_keys, cluster_indices[1:])
+
+        biggest_cluster_index = np.argmax([len(i) for i in clustered_points])
+        # biggest_cluster = clustered_points[biggest_cluster_index]
+        biggest_cluster_keys = clustered_keys[biggest_cluster_index]
+        biggest_cluster_keys = set(biggest_cluster_keys)
+
     for i, (gest, depthcm, K, R, t, imagekey) in enumerate(zip(entry['gestalt'],
                                                                entry['depthcm'],
                                                                entry['K'],
@@ -455,7 +483,11 @@ def predict(entry, visualize=False, scale_estimation_coefficient=2.5, **kwargs)
                 continue
             belonging_points = []
             for i in image_dict[imagekey].point3D_ids[np.where(image_dict[imagekey].point3D_ids != -1)]:
-                belonging_points.append(entry["points3d"][i])
+                if not too_big:
+                    if i in biggest_cluster_keys:
+                        belonging_points.append(entry["points3d"][i])
+                else:
+                    belonging_points.append(entry["points3d"][i])
 
             if len(belonging_points) < 1:
                 print(f'No 3D points in image {i}')

script.py

@@ -132,7 +132,7 @@ if __name__ == "__main__":
                                        point_radius=25,
                                        max_angle=15,
                                        extend=30,
-                                       merge_th=3.0,
+                                       merge_th=100.0,
                                        min_missing_distance=30000000.0,
                                        scale_estimation_coefficient=2.54,
                                        ))