added dbscan

handcrafted_solution.py

@@ -14,6 +14,7 @@ from hoho.read_write_colmap import read_cameras_binary, read_images_binary, read
 from scipy.spatial import KDTree
 from scipy.spatial.distance import cdist
 import scipy.cluster.hierarchy as shc
+from sklearn.cluster import DBSCAN
 
 apex_color = gestalt_color_mapping["apex"]
 eave_end_point = gestalt_color_mapping["eave_end_point"]
@@ -185,6 +186,7 @@ def infer_missing_vertices(ridge_edges, rake_edges):
     missing_candidates = rake_ends.query_ball_tree(ridge_ends, 5)
     missing_candidates = np.concatenate([*missing_candidates])
     missing_candidates = np.unique(missing_candidates).astype(np.int32)
+
     return ridge_ends.data[missing_candidates]
 
 
@@ -439,7 +441,7 @@ def predict(entry, visualize=False, scale_estimation_coefficient=2.5, **kwargs)
     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
+    too_big = False
 
     if not too_big:
         points = np.array(points)
@@ -448,7 +450,7 @@ def predict(entry, visualize=False, scale_estimation_coefficient=2.5, **kwargs)
 
         # print(len(points))
 
-        clustered = shc.fclusterdata(points, 100, criterion='distance')
+        clustered = DBSCAN(eps=100, min_samples=10).fit(points).labels_
         clustered_indices = np.argsort(clustered)
 
         points = points[clustered_indices]