Output OpenCv to virtual camera in Python

Question

I am trying to use OpenCV to detect human faces and draw a rectangle around it, then output it to a virtual camera. I am using pyvirtualcam However, I do not know how to convert the images from the OpenCV VideoCapture to the image format that pyvirtualcam uses. Due to my lack of knowledge, I do not know what format either of them is in, or how to convert one to the other. How do I convert the images from OpenCV to something I can output?

import cv2
import pyvirtualcam
import cvlib as cv


video = cv2.VideoCapture(0)

with pyvirtualcam.Camera(1280, 720, 20) as camera:
    while True:
        ret, im = video.read()

        faces, confidences = cv.detect_face(im)
        for face in faces:
            Rect(face[0], face[1], face[2], face[3]).draw(im)

        camera.send(im)

        camera.sleep_until_next_frame()

import cv2


class Rect:
    def __init__(self, x0, y0, x1, y1):
        self.points = [
            (x0, y0),
            (x1, y1)
        ]
        self.origin = (x0, y0)
        self.width = abs(x1 - x0)
        self.height = abs(y1 - y0)

    def draw(self, im, color=(0, 255, 0), width=3):
        cv2.rectangle(
            im,
            self.points[0],
            self.points[1],
            color,
            width
        )

Answer 1

OpenCV uses BGR. pyvirtualcam by default accepts RGB but also supports BGR and others.

The following should work:

fmt = pyvirtualcam.PixelFormat.BGR
with pyvirtualcam.Camera(1280, 720, 20, fmt=fmt) as camera:

See also the webcam_filter.py sample which is similar to what you try to do.