def test_synthetic_sampson(self, device, dtype):
scene: Dict[str, torch.Tensor] = utils.generate_two_view_random_scene(device, dtype)
x1 = scene['x1']
x2 = scene['x2']
weights = torch.ones_like(x1)[..., 0]
F_est = epi.find_fundamental(x1, x2, weights)
error = epi.sampson_epipolar_distance(x1, x2, F_est)
assert_close(error, torch.tensor(0.0, device=device, dtype=dtype), atol=1e-4, rtol=1e-4)
def test_shift(self, device, dtype):
pts1 = torch.zeros(3, 2, device=device, dtype=dtype)[None]
pts2 = torch.tensor([[2, 0.0], [2, 1], [2, 2.0]],
device=device,
dtype=dtype)[None]
Fm = torch.tensor([[0.0, 0.0, 0.0], [0.0, 0.0, -1.0], [0.0, 1.0, 0.0]],
dtype=dtype,
device=device)[None]
expected = torch.tensor([0.0, 0.5, 2.0], device=device,
dtype=dtype)[None]
assert_close(epi.sampson_epipolar_distance(pts1, pts2, Fm),
expected,
atol=1e-4,
rtol=1e-4)
def test_real_clean(self, device, dtype, data):
torch.random.manual_seed(0)
# generate input data
data_dev = utils.dict_to(data, device, dtype)
pts_src = data_dev['pts0']
pts_dst = data_dev['pts1']
# compute transform from source to target
ransac = RANSAC('fundamental',
inl_th=0.5,
max_iter=20,
max_lo_iters=10).to(device=device, dtype=dtype)
fundamental_matrix, _ = ransac(pts_src, pts_dst)
gross_errors = (sampson_epipolar_distance(pts_src[None],
pts_dst[None],
fundamental_matrix[None],
squared=False) > 1.0)
assert gross_errors.sum().item() == 0
def test_real_dirty(self, device, dtype, data):
torch.random.manual_seed(0)
# generate input data
data_dev = utils.dict_to(data, device, dtype)
pts_src = data_dev['pts0']
pts_dst = data_dev['pts1']
kp1 = data_dev['loftr_indoor_tentatives0']
kp2 = data_dev['loftr_indoor_tentatives1']
ransac = RANSAC('fundamental',
inl_th=1.0,
max_iter=20,
max_lo_iters=10).to(device=device, dtype=dtype)
# compute transform from source to target
fundamental_matrix, _ = ransac(kp1, kp2)
gross_errors = (sampson_epipolar_distance(pts_src[None],
pts_dst[None],
fundamental_matrix[None],
squared=False) > 10.0)
assert gross_errors.sum().item() < 2
def test_batch(self, device, dtype):
batch_size = 5
pts1 = torch.rand(batch_size, 4, 3, device=device, dtype=dtype)
pts2 = torch.rand(batch_size, 4, 3, device=device, dtype=dtype)
Fm = utils.create_random_fundamental_matrix(1).type_as(pts1)
assert epi.sampson_epipolar_distance(pts1, pts2, Fm).shape == (5, 4)
