def test_transform_boxes_wh(self, device, dtype):
boxes = torch.tensor(
[
[139.2640, 103.0150, 258.0480, 307.5075],
[1.0240, 80.5547, 510.9760, 431.4453],
[165.2053, 262.1440, 345.4293, 246.7840],
[119.8080, 144.2067, 137.2160, 265.9225],
],
device=device,
dtype=dtype,
)
expected = torch.tensor(
[
[372.7360, 103.0150, -258.0480, 307.5075],
[510.9760, 80.5547, -510.9760, 431.4453],
[346.7947, 262.1440, -345.4293, 246.7840],
[392.1920, 144.2067, -137.2160, 265.9225],
],
device=device,
dtype=dtype,
)
trans_mat = torch.tensor(
[[[-1.0, 0.0, 512.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0]]],
device=device,
dtype=dtype)
out = kornia.transform_boxes(trans_mat, boxes, mode='xywh')
assert_close(out, expected, atol=1e-4, rtol=1e-4)
def test_transform_multiple_boxes(self, device, dtype):
boxes = torch.tensor([[139.2640, 103.0150, 397.3120, 410.5225],
[1.0240, 80.5547, 512.0000, 512.0000],
[165.2053, 262.1440, 510.6347, 508.9280],
[119.8080, 144.2067, 257.0240, 410.1292]],
device=device,
dtype=dtype)
boxes = boxes.repeat(2, 1, 1) # 2 x 4 x 4 two images 4 boxes each
expected = torch.tensor([[[372.7360, 103.0150, 114.6880, 410.5225],
[510.9760, 80.5547, 0.0000, 512.0000],
[346.7947, 262.1440, 1.3653, 508.9280],
[392.1920, 144.2067, 254.9760, 410.1292]],
[[139.2640, 103.0150, 397.3120, 410.5225],
[1.0240, 80.5547, 512.0000, 512.0000],
[165.2053, 262.1440, 510.6347, 508.9280],
[119.8080, 144.2067, 257.0240, 410.1292]]],
device=device,
dtype=dtype)
trans_mat = torch.tensor(
[[[-1., 0., 512.], [0., 1., 0.], [0., 0., 1.]],
[[1., 0., 0.], [0., 1., 0.], [0., 0., 1.]]],
device=device,
dtype=dtype)
out = kornia.transform_boxes(trans_mat, boxes)
assert_allclose(out, expected, atol=1e-4, rtol=1e-4)
def test_transform_boxes(self, device, dtype):
boxes = torch.tensor([[139.2640, 103.0150, 397.3120, 410.5225]], device=device, dtype=dtype)
expected = torch.tensor([[372.7360, 103.0150, 114.6880, 410.5225]], device=device, dtype=dtype)
trans_mat = torch.tensor([[[-1.0, 0.0, 512.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0]]], device=device, dtype=dtype)
out = kornia.transform_boxes(trans_mat, boxes)
assert_allclose(out, expected, atol=1e-4, rtol=1e-4)
def test_transform_boxes(self, device):
boxes = torch.tensor([[139.2640, 103.0150, 397.3120,
410.5225]]).to(device)
expected = torch.tensor([372.7360, 103.0150, 114.6880,
410.5225]).to(device)
trans_mat = torch.tensor([[[-1., 0., 512.], [0., 1., 0.],
[0., 0., 1.]]]).to(device)
out = kornia.transform_boxes(trans_mat, boxes)
assert_allclose(out, expected)
def op_script(transform, boxes):
return kornia.transform_boxes(transform, boxes)