From 8abd806d7abba0978fc617d16c03e122659dffe5 Mon Sep 17 00:00:00 2001
From: brimoor <brimoor@umich.edu>
Date: Fri, 24 May 2024 18:57:57 -0400
Subject: [PATCH] adding padding option

---
 fiftyone/utils/utils3d.py | 40 +++++++++++++++++++++++++++++----------
 1 file changed, 30 insertions(+), 10 deletions(-)

diff --git a/fiftyone/utils/utils3d.py b/fiftyone/utils/utils3d.py
index 388f092033..4e252bb373 100644
--- a/fiftyone/utils/utils3d.py
+++ b/fiftyone/utils/utils3d.py
@@ -535,6 +535,7 @@ def compute_orthographic_projection_images(
     subsampling_rate=None,
     projection_normal=None,
     bounds=None,
+    padding=None,
     skip_failures=False,
     progress=None,
 ):
@@ -600,6 +601,11 @@ def compute_orthographic_projection_images(
             to generate each map. Either element of the tuple or any/all of its
             values can be None, in which case a tight crop of the point cloud
             along the missing dimension(s) are used
+        padding (None): a relative padding(s) in ``[0, 1]]`` to apply to the
+            field of view bounds prior to projection. Can either be a single
+            value to apply in all directions or a ``[padx, pady, padz]``. For
+            example, pass ``padding=0.05`` with no ``bounds`` to project onto
+            a tight crop of each point cloud with 5% padding around it
         skip_failures (False): whether to gracefully continue without raising
             an error if a projection fails
         progress (None): whether to render a progress bar (True/False), use the
@@ -647,6 +653,7 @@ def compute_orthographic_projection_images(
                 subsampling_rate=subsampling_rate,
                 projection_normal=projection_normal,
                 bounds=bounds,
+                padding=padding,
             )
         except Exception as e:
             if not skip_failures:
@@ -680,6 +687,7 @@ def compute_orthographic_projection_image(
     subsampling_rate=None,
     projection_normal=None,
     bounds=None,
+    padding=None,
 ):
     """Generates an orthographic projection image for the given PCD file onto
     the specified plane (default xy plane).
@@ -714,6 +722,11 @@ def compute_orthographic_projection_image(
             the projected plane. Either element of the tuple or any/all of its
             values can be None, in which case a tight crop of the point cloud
             along the missing dimension(s) are used
+        padding (None): a relative padding(s) in ``[0, 1]]`` to apply to the
+            field of view bounds prior to projection. Can either be a single
+            value to apply in all directions or a ``[padx, pady, padz]``. For
+            example, pass ``padding=0.05`` with no ``bounds`` to project onto
+            a tight crop of the point cloud with 5% padding around it
 
     Returns:
         a tuple of
@@ -731,6 +744,7 @@ def compute_orthographic_projection_image(
         filepath,
         size=size,
         bounds=bounds,
+        padding=padding,
         projection_normal=projection_normal,
         subsampling_rate=subsampling_rate,
     )
@@ -828,6 +842,7 @@ def _parse_point_cloud(
     filepath,
     size=None,
     bounds=None,
+    padding=None,
     projection_normal=None,
     subsampling_rate=None,
 ):
@@ -876,19 +891,24 @@ def _parse_point_cloud(
         _max_bound = pc.get_max_bound()
         max_bound = _fill_none(max_bound, _max_bound)
 
+    min_bound = np.asarray(min_bound, dtype=float)
+    max_bound = np.asarray(max_bound, dtype=float)
+
+    if padding is not None:
+        pad = 0.5 * np.asarray(padding) * (max_bound - min_bound)
+        min_bound -= pad
+        max_bound += pad
+
     # Ensure bbox will not have 0 volume by adding a small value if max_bound
-    #   and min_bound are close to each other
-    delta = np.isclose(
-        np.asarray(max_bound) - np.asarray(min_bound), 0
-    ) * np.repeat(0.000001, 3)
-    max_bound += delta
+    # and min_bound are close to each other
+    max_bound += 1e-6 * np.isclose(max_bound - min_bound, 0)
 
     bbox = o3d.geometry.AxisAlignedBoundingBox(
         min_bound=min_bound, max_bound=max_bound
     )
 
-    # crop bounds and translate so that min bound is at the origin
-    pc = pc.crop(bbox).translate((-min_bound[0], -min_bound[1], -min_bound[2]))
+    # Crop bounds and translate so that min bound is at the origin
+    pc = pc.crop(bbox).translate(-min_bound)
 
     if subsampling_rate is not None and subsampling_rate > 0:
         pc = pc.uniform_down_sample(subsampling_rate)
@@ -902,9 +922,9 @@ def _parse_point_cloud(
         width, height = None, None
 
     metadata = OrthographicProjectionMetadata(
-        min_bound=min_bound,
-        max_bound=max_bound,
-        normal=projection_normal,
+        min_bound=list(min_bound),
+        max_bound=list(max_bound),
+        normal=list(projection_normal),
         width=width,
         height=height,
     )