Add circular padding

test/test_transforms_tensor.py

@@ -206,7 +206,7 @@ def test_color_jitter_all(self, device, channels):
 
 
 @pytest.mark.parametrize("device", cpu_and_cuda())
-@pytest.mark.parametrize("m", ["constant", "edge", "reflect", "symmetric"])
+@pytest.mark.parametrize("m", ["constant", "edge", "reflect", "symmetric", "circular"])
 @pytest.mark.parametrize("mul", [1, -1])
 def test_pad(m, mul, device):
     fill = 127 if m == "constant" else 0
@@ -264,6 +264,7 @@ def test_crop(device):
         {"padding_mode": "constant", "fill": 10},
         {"padding_mode": "edge"},
         {"padding_mode": "reflect"},
+        {"padding_mode": "circular"},
     ],
 )
 @pytest.mark.parametrize("pad_if_needed", [True, False])

torchvision/transforms/_functional_pil.py

@@ -145,7 +145,7 @@ def pad(
     img: Image.Image,
     padding: Union[int, List[int], Tuple[int, ...]],
     fill: Optional[Union[float, List[float], Tuple[float, ...]]] = 0,
-    padding_mode: Literal["constant", "edge", "reflect", "symmetric"] = "constant",
+    padding_mode: Literal["constant", "edge", "reflect", "symmetric", "circular"] = "constant",
 ) -> Image.Image:
 
     if not _is_pil_image(img):
@@ -168,8 +168,8 @@ def pad(
         # Compatibility with `functional_tensor.pad`
         padding = padding[0]
 
-    if padding_mode not in ["constant", "edge", "reflect", "symmetric"]:
-        raise ValueError("Padding mode should be either constant, edge, reflect or symmetric")
+    if padding_mode not in ["constant", "edge", "reflect", "symmetric", "circular"]:
+        raise ValueError("Padding mode should be either constant, edge, reflect, symmetric, or circular")
 
     if padding_mode == "constant":
         opts = _parse_fill(fill, img, name="fill")
@@ -201,6 +201,10 @@ def pad(
 
         pad_left, pad_top, pad_right, pad_bottom = np.maximum(p, 0)
 
+        if padding_mode == "circular":
+            # Compatibility with np.pad
+            padding_mode = "wrap"
+
         if img.mode == "P":
             palette = img.getpalette()
             img = np.asarray(img)

torchvision/transforms/_functional_tensor.py

@@ -398,8 +398,8 @@ def pad(
                 f"Padding must be an int or a 1, 2, or 4 element tuple, not a {len(padding)} element tuple"
             )
 
-    if padding_mode not in ["constant", "edge", "reflect", "symmetric"]:
-        raise ValueError("Padding mode should be either constant, edge, reflect or symmetric")
+    if padding_mode not in ["constant", "edge", "reflect", "symmetric", "circular"]:
+        raise ValueError("Padding mode should be either constant, edge, reflect, symmetric, or circular")
 
     p = _parse_pad_padding(padding)
 
@@ -424,7 +424,7 @@ def pad(
         need_cast = True
         img = img.to(torch.float32)
 
-    if padding_mode in ("reflect", "replicate"):
+    if padding_mode in ("reflect", "replicate", "circular"):
         img = torch_pad(img, p, mode=padding_mode)
     else:
         img = torch_pad(img, p, mode=padding_mode, value=float(fill))

torchvision/transforms/functional.py

@@ -482,7 +482,7 @@ def resize(
 def pad(img: Tensor, padding: List[int], fill: Union[int, float] = 0, padding_mode: str = "constant") -> Tensor:
     r"""Pad the given image on all sides with the given "pad" value.
     If the image is torch Tensor, it is expected
-    to have [..., H, W] shape, where ... means at most 2 leading dimensions for mode reflect and symmetric,
+    to have [..., H, W] shape, where ... means at most 2 leading dimensions for modes reflect, symmetric, and circular,
     at most 3 leading dimensions for mode edge,
     and an arbitrary number of leading dimensions for mode constant
 
@@ -501,7 +501,7 @@ def pad(img: Tensor, padding: List[int], fill: Union[int, float] = 0, padding_mo
             This value is only used when the padding_mode is constant.
             Only number is supported for torch Tensor.
             Only int or tuple value is supported for PIL Image.
-        padding_mode (str): Type of padding. Should be: constant, edge, reflect or symmetric.
+        padding_mode (str): Type of padding. Should be: constant, edge, reflect, symmetric, or circular.
             Default is constant.
 
             - constant: pads with a constant value, this value is specified with fill
@@ -517,6 +517,10 @@ def pad(img: Tensor, padding: List[int], fill: Union[int, float] = 0, padding_mo
               For example, padding [1, 2, 3, 4] with 2 elements on both sides in symmetric mode
               will result in [2, 1, 1, 2, 3, 4, 4, 3]
 
+            - circular: pads by repeating the values from the opposite side of the image in order.
+              For example, padding [1, 2, 3, 4] with 2 elements on both sides in circular mode
+              will result in [3, 4, 1, 2, 3, 4, 1, 2]
+
     Returns:
         PIL Image or Tensor: Padded image.
     """