aertb.core.processing

aertb.core.processing.clean(events, tau=0.5, val=10, R=3, allow_first=200)

Removes some of the noise in an event based scenario

Parameters

  • events (the event structured numpy array) –

  • tau (float, optional) – regulates the decay of the memory, by default 0.5 important to set it properly!

  • val (int, optional) – Mask value threshold, by default 10

  • R (int, optional) – Dimension of the neighborhood, by default 3

  • allow_first (int, optional) – how many events are let through at the beginning without seeing the mask value, by default first 200

Returns

the input structured array without the noise events

Return type

np.array

aertb.core.processing.downscale(events, factor=2, camera_size=None)

Downscales the image by the given factor.

Parameters

  • events (np.array) – structured array with fields (x,y)

  • factor (int, optional) – the downscale factor must perfectly divide the height and with of the image, otherwise it will raise an error., by default 2

  • camera_size (tuple, optional) – Camera size will be assumed from events, if different, it must be specified with this parameter as (width, height), by default None

Returns

the input structured array modified so that the resulting image from events is downscaled by the given factor.

Return type

np.array

aertb.core.processing.flip_diagonal(events)

Modifes a set of events so that the resulting image is flipped along the main diagonal

aertb.core.processing.flip_horizontal(events)

Modifies a set of events so that the resulting image is flipped horizontally

aertb.core.processing.flip_vertical(events)

Modifies a set of events so that the resulting image is flipped vertically

aertb.core.processing.rotate(events, angle, direction='ccw')

Modifes a set of events so that the resulting image is rotated

Parameters

  • events (np.array,) – the input event structured array

  • angle (int,) – degrees to rotate: 90, 180, or 270

  • direction (str, optional) – ‘ccw’ counterclockwise or ‘cw’ clockwise, by default ‘ccw’

Returns

the input array modified so the rotation takes place

Return type

np.array