Circular Mask

[GaGunawan]

Is your feature request related to a problem? Please describe.
I'm currently using the CenturyArk SilkyEvCam and the effective data is in the shape of a circle. If I use the current mask/crop feature, it would either leave unwanted noise, or get rid of wanted data

Describe the solution you'd like
Be able to create a circular mask with variable radius and origin

Additional context
Binarized data to show how a circle is formed in the data

[AdamDHines]

Thanks for opening this issue @GaGunawan.

I'm not sure this is possible as a cropping implementation, a crop will always be a square since images can only be squares - even if cropping with a circular shape. Even OpenCV requires this as a workaround, e.g. -

import cv2
import numpy as np

img = cv2.imread('image.jpg')

# Define your circle parameters
cx, cy, r = 300, 300, 100 

# Crop the bounding square around the circle
crop = img[cy-r : cy+r, cx-r : cx+r]

The principle for this would be the same for EventCV's cropping, e.g. -

import numpy as np
import eventcv as ecv

stream = ecv.load("recording.h5")     # or ecv.from_numpy(...)

cx, cy, r = 300, 300, 100
crop = stream.crop(cx - r, cy - r, 2 * r, 2 * r)   # (x0, y0, width, height)

Does that solve your problem?