Refactor depth reference handling and cleanup

- Unified reference depth calculation to use a single implementation in common/depth.py.
- Removed server-side reference depth computation and adjusted depth.json to only include image_size and intrinsics.
- Updated depth texture handling to ignore transparent, invalid, and non-positive pixels when calculating reference depth.
- Changed reference depth calculation to use the 95th percentile of valid depth values.
- Adjusted depth direction fitting to return a flat direction when no samples are available.
- Removed cleanup threshold parameters from Cutout and related nodes, simplifying the interface.
- Updated tests to verify new reference depth logic and ensure compatibility with existing functionality.

tests/anyimage/operators/test_cutout_boundary_padding.py

@@ -128,8 +128,8 @@ def test_client_processing_updates_and_packs_color_and_depth_together():
     depth = Image(depth_values)
     metadata = {
         "intrinsics": ((8.0, 0.0, 8.0), (0.0, 8.0, 8.0), (0.0, 0.0, 1.0)),
-        "reference_depth": 12.0,
     }
+    original_metadata = dict(metadata)
 
     pad_cutout_images(color, depth, metadata, content, 0.9, 2)
 
@@ -139,5 +139,5 @@ def test_client_processing_updates_and_packs_color_and_depth_together():
     assert padded_color[2, 8, 3] == 1
     assert padded_depth[2, 8, 2] > depth_values[3, 8, 2]
     assert padded_depth[2, 8, 0] == 0
-    assert metadata["reference_depth"] == 12
+    assert metadata == original_metadata
     assert (color.updated, color.packed, depth.updated, depth.packed) == (1, 1, 1, 1)