[fbsync] Added CIOU loss function (#5776)

Summary:
* added ciou loss

* "formatting with flake8 and ufmt"

* formatting with ufmt and flake8

* minor changes

* changes as per the suggestions

* added reference in torchvision/ops/__init__.py

* sample test

* tests formatted

* added description

* formatting

* edited tests

* changes in tests

* added tests for multiple boxes

* minor edits

* minor edit

* doc added

* minor edits

* Update test_ops.py

* formatting test file

* changes as per the suggestions

* formatting and adding some more tests

* bounding box added

* removed unnecessary comment

* added docstring

* added type annotations

* removed potential bug

* Update torchvision/ops/boxes.py

* Update torchvision/ops/boxes.py

* Update test/test_ops.py

Reviewed By: jdsgomes, NicolasHug

Differential Revision: D36095722

fbshipit-source-id: be54569f72c54380a832e0798b83a0cfe6558c86

Co-authored-by: Vasilis Vryniotis <[email protected]>
Co-authored-by: Vasilis Vryniotis <[email protected]>
Co-authored-by: Vasilis Vryniotis <[email protected]>
Co-authored-by: Philip Meier <[email protected]>

Author: 3 people

docs/source/ops.rst

@@ -19,6 +19,8 @@ Operators
     box_convert
     box_iou
     clip_boxes_to_image
+    complete_box_iou
+    complete_box_iou_loss
     deform_conv2d
     drop_block2d
     drop_block3d

test/test_ops.py

@@ -1258,6 +1258,43 @@ def test_giou_jit(self) -> None:
         self._run_jit_test([[0, 0, 100, 100], [0, 0, 50, 50], [200, 200, 300, 300]])
 
 
+class TestCompleteBoxIou(BoxTestBase):
+    def _target_fn(self) -> Tuple[bool, Callable]:
+        return (True, ops.complete_box_iou)
+
+    def _generate_int_input() -> List[List[int]]:
+        return [[0, 0, 100, 100], [0, 0, 50, 50], [200, 200, 300, 300]]
+
+    def _generate_int_expected() -> List[List[float]]:
+        return [[1.0, 0.25, 0.0], [0.25, 1.0, 0.0], [0.0, 0.0, 1.0]]
+
+    def _generate_float_input() -> List[List[float]]:
+        return [
+            [285.3538, 185.5758, 1193.5110, 851.4551],
+            [285.1472, 188.7374, 1192.4984, 851.0669],
+            [279.2440, 197.9812, 1189.4746, 849.2019],
+        ]
+
+    def _generate_float_expected() -> List[List[float]]:
+        return [[1.0, 0.9933, 0.9673], [0.9933, 1.0, 0.9737], [0.9673, 0.9737, 1.0]]
+
+    @pytest.mark.parametrize(
+        "test_input, dtypes, tolerance, expected",
+        [
+            pytest.param(
+                _generate_int_input(), [torch.int16, torch.int32, torch.int64], 1e-4, _generate_int_expected()
+            ),
+            pytest.param(_generate_float_input(), [torch.float32, torch.float64], 0.002, _generate_float_expected()),
+            pytest.param(_generate_float_input(), [torch.float32, torch.float64], 0.001, _generate_float_expected()),
+        ],
+    )
+    def test_complete_iou(self, test_input: List, dtypes: List[torch.dtype], tolerance: float, expected: List) -> None:
+        self._run_test(test_input, dtypes, tolerance, expected)
+
+    def test_ciou_jit(self) -> None:
+        self._run_jit_test([[0, 0, 100, 100], [0, 0, 50, 50], [200, 200, 300, 300]])
+
+
 class TestMasksToBoxes:
     def test_masks_box(self):
         def masks_box_check(masks, expected, tolerance=1e-4):
@@ -1578,6 +1615,7 @@ def test_giou_loss(self, dtype, device) -> None:
         box2 = torch.tensor([0, 0, 1, 1], dtype=dtype, device=device)
         box3 = torch.tensor([0, 1, 1, 2], dtype=dtype, device=device)
         box4 = torch.tensor([1, 1, 2, 2], dtype=dtype, device=device)
+
         box1s = torch.stack([box2, box2], dim=0)
         box2s = torch.stack([box3, box4], dim=0)
 
@@ -1623,5 +1661,53 @@ def test_empty_inputs(self, dtype, device) -> None:
         assert loss.numel() == 0, "giou_loss for two empty box should be empty"
 
 
+class TestCIOULoss:
+    @pytest.mark.parametrize("dtype", [torch.float32, torch.half])
+    @pytest.mark.parametrize("device", cpu_and_gpu())
+    def test_ciou_loss(self, dtype, device):
+        box1 = torch.tensor([-1, -1, 1, 1], dtype=dtype, device=device)
+        box2 = torch.tensor([0, 0, 1, 1], dtype=dtype, device=device)
+        box3 = torch.tensor([0, 1, 1, 2], dtype=dtype, device=device)
+        box4 = torch.tensor([1, 1, 2, 2], dtype=dtype, device=device)
+
+        box1s = torch.stack([box2, box2], dim=0)
+        box2s = torch.stack([box3, box4], dim=0)
+
+        def assert_ciou_loss(box1, box2, expected_output, reduction="none"):
+
+            output = ops.complete_box_iou_loss(box1, box2, reduction=reduction)
+            expected_output = torch.tensor(expected_output, device=device)
+            tol = 1e-5 if dtype != torch.half else 1e-3
+            torch.testing.assert_close(output, expected_output, rtol=tol, atol=tol)
+
+        assert_ciou_loss(box1, box1, 0.0)
+
+        assert_ciou_loss(box1, box2, 0.8125)
+
+        assert_ciou_loss(box1, box3, 1.1923)
+
+        assert_ciou_loss(box1, box4, 1.2500)
+
+        assert_ciou_loss(box1s, box2s, 1.2250, reduction="mean")
+        assert_ciou_loss(box1s, box2s, 2.4500, reduction="sum")
+
+    @pytest.mark.parametrize("device", cpu_and_gpu())
+    @pytest.mark.parametrize("dtype", [torch.float32, torch.half])
+    def test_empty_inputs(self, dtype, device) -> None:
+        box1 = torch.randn([0, 4], dtype=dtype).requires_grad_()
+        box2 = torch.randn([0, 4], dtype=dtype).requires_grad_()
+
+        loss = ops.complete_box_iou_loss(box1, box2, reduction="mean")
+        loss.backward()
+
+        tol = 1e-3 if dtype is torch.half else 1e-5
+        torch.testing.assert_close(loss, torch.tensor(0.0), rtol=tol, atol=tol)
+        assert box1.grad is not None, "box1.grad should not be None after backward is called"
+        assert box2.grad is not None, "box2.grad should not be None after backward is called"
+
+        loss = ops.complete_box_iou_loss(box1, box2, reduction="none")
+        assert loss.numel() == 0, "ciou_loss for two empty box should be empty"
+
+
 if __name__ == "__main__":
     pytest.main([__file__])

torchvision/ops/__init__.py

@@ -7,9 +7,11 @@
     box_area,
     box_iou,
     generalized_box_iou,
+    complete_box_iou,
     masks_to_boxes,
 )
 from .boxes import box_convert
+from .ciou_loss import complete_box_iou_loss
 from .deform_conv import deform_conv2d, DeformConv2d
 from .drop_block import drop_block2d, DropBlock2d, drop_block3d, DropBlock3d
 from .feature_pyramid_network import FeaturePyramidNetwork

torchvision/ops/boxes.py

@@ -311,6 +311,54 @@ def generalized_box_iou(boxes1: Tensor, boxes2: Tensor) -> Tensor:
     return iou - (areai - union) / areai
 
 
+def complete_box_iou(boxes1: Tensor, boxes2: Tensor, eps: float = 1e-7) -> Tensor:
+    """
+    Return complete intersection-over-union (Jaccard index) between two sets of boxes.
+    Both sets of boxes are expected to be in ``(x1, y1, x2, y2)`` format with
+    ``0 <= x1 < x2`` and ``0 <= y1 < y2``.
+    Args:
+        boxes1 (Tensor[N, 4]): first set of boxes
+        boxes2 (Tensor[M, 4]): second set of boxes
+        eps (float, optional): small number to prevent division by zero. Default: 1e-7
+    Returns:
+        Tensor[N, M]: the NxM matrix containing the pairwise complete IoU values
+        for every element in boxes1 and boxes2
+    """
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(complete_box_iou)
+
+    boxes1 = _upcast(boxes1)
+    boxes2 = _upcast(boxes2)
+
+    inter, union = _box_inter_union(boxes1, boxes2)
+    iou = inter / union
+
+    lti = torch.min(boxes1[:, None, :2], boxes2[:, None, :2])
+    rbi = torch.max(boxes1[:, None, 2:], boxes2[:, None, 2:])
+
+    whi = (rbi - lti).clamp(min=0)  # [N,M,2]
+    diagonal_distance_squared = (whi[:, :, 0] ** 2) + (whi[:, :, 1] ** 2) + eps
+
+    # centers of boxes
+    x_p = (boxes1[:, 0] + boxes1[:, 2]) / 2
+    y_p = (boxes1[:, 1] + boxes1[:, 3]) / 2
+    x_g = (boxes2[:, 0] + boxes2[:, 2]) / 2
+    y_g = (boxes2[:, 1] + boxes2[:, 3]) / 2
+    # The distance between boxes' centers squared.
+    centers_distance_squared = (x_p - x_g) ** 2 + (y_p - y_g) ** 2
+
+    w_pred = boxes1[:, 2] - boxes1[:, 0]
+    h_pred = boxes1[:, 3] - boxes1[:, 1]
+
+    w_gt = boxes2[:, 2] - boxes2[:, 0]
+    h_gt = boxes2[:, 3] - boxes2[:, 1]
+
+    v = (4 / (torch.pi ** 2)) * torch.pow((torch.atan(w_gt / h_gt) - torch.atan(w_pred / h_pred)), 2)
+    with torch.no_grad():
+        alpha = v / (1 - iou + v + eps)
+    return iou - (centers_distance_squared / diagonal_distance_squared) - alpha * v
+
+
 def masks_to_boxes(masks: torch.Tensor) -> torch.Tensor:
     """
     Compute the bounding boxes around the provided masks.

torchvision/ops/ciou_loss.py

@@ -0,0 +1,92 @@
+import torch
+
+from ..utils import _log_api_usage_once
+from .giou_loss import _upcast
+
+
+def complete_box_iou_loss(
+    boxes1: torch.Tensor,
+    boxes2: torch.Tensor,
+    reduction: str = "none",
+    eps: float = 1e-7,
+) -> torch.Tensor:
+
+    """
+    Gradient-friendly IoU loss with an additional penalty that is non-zero when the
+    boxes do not overlap overlap area, This loss function considers important geometrical
+    factors such as  overlap area, normalized central point distance and aspect ratio.
+    This loss is symmetric, so the boxes1 and boxes2 arguments are interchangeable.
+
+    Both sets of boxes are expected to be in ``(x1, y1, x2, y2)`` format with
+    ``0 <= x1 < x2`` and ``0 <= y1 < y2``, and The two boxes should have the
+    same dimensions.
+
+    Args:
+        boxes1 : (Tensor[N, 4] or Tensor[4]) first set of boxes
+        boxes2 : (Tensor[N, 4] or Tensor[4]) second set of boxes
+        reduction : (string, optional) Specifies the reduction to apply to the output:
+            ``'none'`` | ``'mean'`` | ``'sum'``. ``'none'``: No reduction will be
+            applied to the output. ``'mean'``: The output will be averaged.
+            ``'sum'``: The output will be summed. Default: ``'none'``
+        eps : (float): small number to prevent division by zero. Default: 1e-7
+
+    Reference:
+
+    Complete Intersection over Union Loss (Zhaohui Zheng et. al)
+    https://arxiv.org/abs/1911.08287
+
+    """
+
+    # Original Implementation : https://github.com/facebookresearch/detectron2/blob/main/detectron2/layers/losses.py
+
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(complete_box_iou_loss)
+
+    boxes1 = _upcast(boxes1)
+    boxes2 = _upcast(boxes2)
+
+    x1, y1, x2, y2 = boxes1.unbind(dim=-1)
+    x1g, y1g, x2g, y2g = boxes2.unbind(dim=-1)
+
+    # Intersection keypoints
+    xkis1 = torch.max(x1, x1g)
+    ykis1 = torch.max(y1, y1g)
+    xkis2 = torch.min(x2, x2g)
+    ykis2 = torch.min(y2, y2g)
+
+    intsct = torch.zeros_like(x1)
+    mask = (ykis2 > ykis1) & (xkis2 > xkis1)
+    intsct[mask] = (xkis2[mask] - xkis1[mask]) * (ykis2[mask] - ykis1[mask])
+    union = (x2 - x1) * (y2 - y1) + (x2g - x1g) * (y2g - y1g) - intsct + eps
+    iou = intsct / union
+
+    # smallest enclosing box
+    xc1 = torch.min(x1, x1g)
+    yc1 = torch.min(y1, y1g)
+    xc2 = torch.max(x2, x2g)
+    yc2 = torch.max(y2, y2g)
+    diag_len = ((xc2 - xc1) ** 2) + ((yc2 - yc1) ** 2) + eps
+
+    # centers of boxes
+    x_p = (x2 + x1) / 2
+    y_p = (y2 + y1) / 2
+    x_g = (x1g + x2g) / 2
+    y_g = (y1g + y2g) / 2
+    distance = ((x_p - x_g) ** 2) + ((y_p - y_g) ** 2)
+
+    # width and height of boxes
+    w_pred = x2 - x1
+    h_pred = y2 - y1
+    w_gt = x2g - x1g
+    h_gt = y2g - y1g
+    v = (4 / (torch.pi ** 2)) * torch.pow((torch.atan(w_gt / h_gt) - torch.atan(w_pred / h_pred)), 2)
+    with torch.no_grad():
+        alpha = v / (1 - iou + v + eps)
+
+    loss = 1 - iou + (distance / diag_len) + alpha * v
+    if reduction == "mean":
+        loss = loss.mean() if loss.numel() > 0 else 0.0 * loss.sum()
+    elif reduction == "sum":
+        loss = loss.sum()
+
+    return loss