def test_representation_voxel_grid(self, n_time_bins):
orig_events, sensor_size = create_random_input()
transform = transforms.ToVoxelGrid(
sensor_size=sensor_size, n_time_bins=n_time_bins
)
volume = transform(orig_events)
assert volume.shape == (n_time_bins, *sensor_size[1::-1])
assert volume is not orig_events
def test_representation_frame(
self,
time_window,
event_count,
n_time_bins,
n_event_bins,
overlap,
include_incomplete,
sensor_size,
):
orig_events, _ = create_random_input(sensor_size=sensor_size)
transform = transforms.ToFrame(
sensor_size=sensor_size,
time_window=time_window,
event_count=event_count,
n_time_bins=n_time_bins,
n_event_bins=n_event_bins,
overlap=overlap,
include_incomplete=include_incomplete,
)
frames = transform(orig_events)
assert frames.shape[1:] == sensor_size[::-1]
if time_window is not None:
stride = time_window - overlap
times = orig_events["t"]
if include_incomplete:
assert frames.shape[0] == int(
np.ceil(((times[-1] - times[0]) - time_window) / stride) + 1
)
else:
assert frames.shape[0] == int(
np.floor(((times[-1] - times[0]) - time_window) / stride) + 1
)
if event_count is not None:
stride = event_count - overlap
n_events = orig_events.shape[0]
if include_incomplete:
assert frames.shape[0] == int(
np.ceil((n_events - event_count) / stride) + 1
)
else:
assert frames.shape[0] == int(
np.floor((n_events - event_count) / stride) + 1
)
if n_time_bins is not None:
assert frames.shape[0] == n_time_bins
if n_event_bins is not None:
assert frames.shape[0] == n_event_bins
assert frames is not orig_events
def test_representation_avg_time_surface(self, surface_size, tau):
orig_events, sensor_size = create_random_input()
transform = transforms.ToAveragedTimesurface(
sensor_size=sensor_size, surface_size=surface_size, tau=tau
)
surfaces = transform(orig_events)
assert surfaces.shape[0] == len(orig_events)
assert surfaces.shape[1] == 2
assert surfaces.shape[2] == surface_size
assert surfaces.shape[3] == surface_size
assert surfaces is not orig_events
def test_representation_time_surface(self, surface_dimensions, tau):
orig_events, sensor_size = create_random_input()
transform = transforms.ToTimesurface(
sensor_size=sensor_size, surface_dimensions=surface_dimensions, tau=tau
)
surfaces = transform(orig_events)
assert surfaces.shape[0] == len(orig_events)
assert surfaces.shape[1] == 2
if surface_dimensions:
assert surfaces.shape[2:] == surface_dimensions
else:
assert surfaces.shape[2] == sensor_size[1]
assert surfaces.shape[3] == sensor_size[0]
assert surfaces is not orig_events
def test_representation_frame_inferred(self):
sensor_size = (20, 10, 2)
orig_events, _ = create_random_input(n_events=30000, sensor_size=sensor_size)
transform = transforms.ToFrame(sensor_size=None, time_window=25000)
frames = transform(orig_events)
assert frames.shape[1:] == sensor_size[::-1]
