new script to generate edge topology

share/scripts/generate-edge-topology.py

@@ -0,0 +1,46 @@
+# This script computes an edge_topology.json file for a given model, which is
+# used in eos's contour fitting. It can be used for any Morphable Model,
+# for example the SFM, BFM2009, BFM2017, and others.
+# The general idea of this code has been taken from:
+# https://github.com/waps101/3DMM_edges:
+# A. Bas, W.A.P. Smith, T. Bolkart and S. Wuhrer,
+# "Fitting a 3D Morphable Model to Edges: A Comparison Between Hard and Soft
+#  Correspondences", ACCV Workshop 2016.
+
+
+# Set input and output paths for the model that you want to process:
+MODEL_PATH = "sfm_shape_3448.bin"
+JSON_PATH = "bfm2017-1_bfm_nomouth_edge_topology.json"
+
+
+def generate_edge_topology(triangles):
+    print("triangles:", len(triangles))
+    edges = []
+    indices = []
+
+    for idx, (v0, v1, v2) in enumerate(triangles, start=1):
+        # add 1 to all indices to make 1-based indices for Matlab
+        v0, v1, v2 = sorted([v0+1, v1+1, v2+1])
+        for e in [[v0, v1], [v0, v2], [v1, v2]]:
+            try:
+                i = edges.index(e)
+            except ValueError:
+                edges.append(e)
+                indices.append([0, idx])
+            else:
+                indices[i] = [indices[i][1], idx]
+    print("edges:", len(edges), len(indices))
+    return indices, edges
+
+
+if __name__ == "__main__":
+    import eos
+
+    model = eos.morphablemodel.load_model(MODEL_PATH)
+    triangles = model.get_shape_model().get_triangle_list()
+    indices, edges = generate_edge_topology(triangles)
+
+    # Save it as an eos EdgeTopology in json format:
+    edge_topology = eos.morphablemodel.EdgeTopology(indices, edges)
+    eos.morphablemodel.save_edge_topology(edge_topology, JSON_PATH)
+    print("Finished generating edge topology file and saved it as", JSON_PATH)