[fbsync] Avoid recommuting the affine matrix in bbox rotate (#6712)

Summary:
* Avoid recommuting the affine matrix in bbox rotate

* Fix linter

* inverted=True for estimating image size

* Update the image size estimation to match the one from the image kernel

* Nits

* Address comments.

* Center=0,0 when expand=true

Reviewed By: NicolasHug

Differential Revision: D40427463

fbshipit-source-id: 87e147e8a52c18d90601c6de4b1b182c60b9d8e9

Author: Yosua Michael Maranatha

torchvision/prototype/transforms/functional/_geometry.py

@@ -279,9 +279,9 @@ def affine_image_tensor(
     center_f = [0.0, 0.0]
     if center is not None:
         # Center values should be in pixel coordinates but translated such that (0, 0) corresponds to image center.
-        center_f = [1.0 * (c - s * 0.5) for c, s in zip(center, [width, height])]
+        center_f = [(c - s * 0.5) for c, s in zip(center, [width, height])]
 
-    translate_f = [1.0 * t for t in translate]
+    translate_f = [float(t) for t in translate]
     matrix = _get_inverse_affine_matrix(center_f, angle, translate_f, scale, shear)
 
     output = _FT.affine(image, matrix, interpolation=interpolation.value, fill=fill)
@@ -321,7 +321,7 @@ def _affine_bounding_box_xyxy(
     shear: List[float],
     center: Optional[List[float]] = None,
     expand: bool = False,
-) -> torch.Tensor:
+) -> Tuple[torch.Tensor, Tuple[int, int]]:
     angle, translate, shear, center = _affine_parse_args(
         angle, translate, scale, shear, InterpolationMode.NEAREST, center
     )
@@ -333,19 +333,24 @@ def _affine_bounding_box_xyxy(
     dtype = bounding_box.dtype if torch.is_floating_point(bounding_box) else torch.float32
     device = bounding_box.device
 
-    affine_matrix = torch.tensor(
-        _get_inverse_affine_matrix(center, angle, translate, scale, shear, inverted=False),
-        dtype=dtype,
-        device=device,
-    ).view(2, 3)
+    affine_vector = _get_inverse_affine_matrix(center, angle, translate, scale, shear, inverted=False)
+    transposed_affine_matrix = (
+        torch.tensor(
+            affine_vector,
+            dtype=dtype,
+            device=device,
+        )
+        .view(2, 3)
+        .T
+    )
     # 1) Let's transform bboxes into a tensor of 4 points (top-left, top-right, bottom-left, bottom-right corners).
     # Tensor of points has shape (N * 4, 3), where N is the number of bboxes
     # Single point structure is similar to
     # [(xmin, ymin, 1), (xmax, ymin, 1), (xmax, ymax, 1), (xmin, ymax, 1)]
     points = bounding_box[:, [[0, 1], [2, 1], [2, 3], [0, 3]]].view(-1, 2)
     points = torch.cat([points, torch.ones(points.shape[0], 1, device=points.device)], dim=-1)
     # 2) Now let's transform the points using affine matrix
-    transformed_points = torch.matmul(points, affine_matrix.T)
+    transformed_points = torch.matmul(points, transposed_affine_matrix)
     # 3) Reshape transformed points to [N boxes, 4 points, x/y coords]
     # and compute bounding box from 4 transformed points:
     transformed_points = transformed_points.view(-1, 4, 2)
@@ -360,20 +365,24 @@ def _affine_bounding_box_xyxy(
         points = torch.tensor(
             [
                 [0.0, 0.0, 1.0],
-                [0.0, 1.0 * height, 1.0],
-                [1.0 * width, 1.0 * height, 1.0],
-                [1.0 * width, 0.0, 1.0],
+                [0.0, float(height), 1.0],
+                [float(width), float(height), 1.0],
+                [float(width), 0.0, 1.0],
             ],
             dtype=dtype,
             device=device,
         )
-        new_points = torch.matmul(points, affine_matrix.T)
+        new_points = torch.matmul(points, transposed_affine_matrix)
         tr, _ = torch.min(new_points, dim=0, keepdim=True)
         # Translate bounding boxes
         out_bboxes[:, 0::2] = out_bboxes[:, 0::2] - tr[:, 0]
         out_bboxes[:, 1::2] = out_bboxes[:, 1::2] - tr[:, 1]
+        # Estimate meta-data for image with inverted=True and with center=[0,0]
+        affine_vector = _get_inverse_affine_matrix([0.0, 0.0], angle, translate, scale, shear)
+        new_width, new_height = _FT._compute_affine_output_size(affine_vector, width, height)
+        image_size = (new_height, new_width)
 
-    return out_bboxes.to(bounding_box.dtype)
+    return out_bboxes.to(bounding_box.dtype), image_size
 
 
 def affine_bounding_box(
@@ -391,7 +400,7 @@ def affine_bounding_box(
         bounding_box, old_format=format, new_format=features.BoundingBoxFormat.XYXY
     ).view(-1, 4)
 
-    out_bboxes = _affine_bounding_box_xyxy(bounding_box, image_size, angle, translate, scale, shear, center)
+    out_bboxes, _ = _affine_bounding_box_xyxy(bounding_box, image_size, angle, translate, scale, shear, center)
 
     # out_bboxes should be of shape [N boxes, 4]
 
@@ -502,7 +511,7 @@ def rotate_image_tensor(
             warnings.warn("The provided center argument has no effect on the result if expand is True")
         else:
             # Center values should be in pixel coordinates but translated such that (0, 0) corresponds to image center.
-            center_f = [1.0 * (c - s * 0.5) for c, s in zip(center, [width, height])]
+            center_f = [(c - s * 0.5) for c, s in zip(center, [width, height])]
 
     # due to current incoherence of rotation angle direction between affine and rotate implementations
     # we need to set -angle.
@@ -558,7 +567,7 @@ def rotate_bounding_box(
         bounding_box, old_format=format, new_format=features.BoundingBoxFormat.XYXY
     ).view(-1, 4)
 
-    out_bboxes = _affine_bounding_box_xyxy(
+    out_bboxes, image_size = _affine_bounding_box_xyxy(
         bounding_box,
         image_size,
         angle=-angle,
@@ -569,14 +578,6 @@ def rotate_bounding_box(
         expand=expand,
     )
 
-    if expand:
-        # TODO: Move this computation inside of `_affine_bounding_box_xyxy` to avoid computing the rotation and points
-        #  matrix twice
-        height, width = image_size
-        rotation_matrix = _get_inverse_affine_matrix([0.0, 0.0], angle, [0.0, 0.0], 1.0, [0.0, 0.0])
-        new_width, new_height = _FT._compute_affine_output_size(rotation_matrix, width, height)
-        image_size = (new_height, new_width)
-
     return (
         convert_format_bounding_box(
             out_bboxes, old_format=features.BoundingBoxFormat.XYXY, new_format=format, copy=False