Adding a Checkerboard mesh utility to Pytorch3d

Summary: Adding a checkerboard mesh utility to Pytorch3d.

Reviewed By: bottler

Differential Revision: D39718916

fbshipit-source-id: d43cd30e566b5db068bae6eed0388057634428c8

pytorch3d/utils/__init__.py

@@ -10,6 +10,7 @@ from .camera_conversions import (
     pulsar_from_cameras_projection,
     pulsar_from_opencv_projection,
 )
+from .checkerboard import checkerboard
 from .ico_sphere import ico_sphere
 from .torus import torus
 

pytorch3d/utils/checkerboard.py

@@ -0,0 +1,89 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+
+from typing import Optional, Tuple
+
+import torch
+from pytorch3d.common.compat import meshgrid_ij
+from pytorch3d.renderer.mesh.textures import TexturesAtlas
+from pytorch3d.structures.meshes import Meshes
+
+
+def checkerboard(
+    radius: int = 4,
+    color1: Tuple[float, ...] = (0.0, 0.0, 0.0),
+    color2: Tuple[float, ...] = (1.0, 1.0, 1.0),
+    device: Optional[torch.types._device] = None,
+) -> Meshes:
+    """
+    Returns a mesh of squares in the xy-plane where each unit is one of the two given
+    colors and adjacent squares have opposite colors.
+    Args:
+        radius: how many squares in each direction from the origin
+        color1: background color
+        color2: foreground color (must have the same number of channels as color1)
+    Returns:
+        new Meshes object containing one mesh.
+    """
+
+    if device is None:
+        device = torch.device("cpu")
+    if radius < 1:
+        raise ValueError("radius must be > 0")
+
+    num_verts_per_row = 2 * radius + 1
+
+    # construct 2D grid of 3D vertices
+    x = torch.arange(-radius, radius + 1, device=device)
+    grid_y, grid_x = meshgrid_ij(x, x)
+    verts = torch.stack(
+        [grid_x, grid_y, torch.zeros((2 * radius + 1, 2 * radius + 1))], dim=-1
+    )
+    verts = verts.view(1, -1, 3)
+
+    top_triangle_idx = torch.arange(0, num_verts_per_row * (num_verts_per_row - 1))
+    top_triangle_idx = torch.stack(
+        [
+            top_triangle_idx,
+            top_triangle_idx + 1,
+            top_triangle_idx + num_verts_per_row + 1,
+        ],
+        dim=-1,
+    )
+
+    bottom_triangle_idx = top_triangle_idx[:, [0, 2, 1]] + torch.tensor(
+        [0, 0, num_verts_per_row - 1]
+    )
+
+    faces = torch.zeros(
+        (1, len(top_triangle_idx) + len(bottom_triangle_idx), 3),
+        dtype=torch.long,
+        device=device,
+    )
+    faces[0, ::2] = top_triangle_idx
+    faces[0, 1::2] = bottom_triangle_idx
+
+    # construct range of indices that excludes the boundary to avoid wrong triangles
+    indexing_range = torch.arange(0, 2 * num_verts_per_row * num_verts_per_row).view(
+        num_verts_per_row, num_verts_per_row, 2
+    )
+    indexing_range = indexing_range[:-1, :-1]  # removes boundaries from list of indices
+    indexing_range = indexing_range.reshape(
+        2 * (num_verts_per_row - 1) * (num_verts_per_row - 1)
+    )
+
+    faces = faces[:, indexing_range]
+
+    # adding color
+    colors = torch.tensor(color1).repeat(2 * num_verts_per_row * num_verts_per_row, 1)
+    colors[2::4] = torch.tensor(color2)
+    colors[3::4] = torch.tensor(color2)
+    colors = colors[None, indexing_range, None, None]
+
+    texture_atlas = TexturesAtlas(colors)
+
+    return Meshes(verts=verts, faces=faces, textures=texture_atlas)

tests/test_checkerboard.py

@@ -0,0 +1,21 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+import unittest
+
+import torch
+from pytorch3d.utils import checkerboard
+
+from .common_testing import TestCaseMixin
+
+
+class TestCheckerboard(TestCaseMixin, unittest.TestCase):
+    def test_simple(self):
+        board = checkerboard(5)
+        verts = board.verts_packed()
+        expect = torch.tensor([5.0, 5.0, 0])
+        self.assertClose(verts.min(dim=0).values, -expect)
+        self.assertClose(verts.max(dim=0).values, expect)